Cortisol Dynamics in the Acute Phase of Aneurysmal Subarachnoid Hemorrhage: Associations with Disease Severity and Outcome

Abstract

The purpose of this study was to assess cortisol dynamics in the acute phase after aneurysmal subarachnoid hemorrhage (SAH) and to set the parameters of cortisol release in relation to the severity of illness and outcome. In 22 consecutive patients with aneurysmal SAH, cortisol, corticosteroid binding globulin, interleukin-6, and adrenocorticotrophic hormone were measured immediately after hospital admission (t ₀), 7 days (t ₁) later, and at least 14 days later (t ₂). Additionally, diurnal profiles of cortisol secretion were assessed at t ₁ and t ₂, and area under the curve (AUC) was computed for calculated free serum cortisol (CFSC). In this study, normal diurnal CFSC profiles were associated with a significantly shorter ICU-stay, less complications, and a more favorable outcome than abnormal diurnal profiles. AUC and 8 a.m. cortisol were not related to clinical course or outcome. It is concluded that cortisol secretion patterns are associated with the severity and outcome of SAH. For an appraisal of the hypothalamo–pituitary–adrenal axis in SAH patients, single cortisol measurements are insufficient.

Introduction

A ctivation of the hypothalamo–pituitary–adrenal axis (HPAA) is an important mechanism in the response to stressful events and critical illness. Not surprisingly, a vast number of studies has given evidence of excessive HPAA activation in patients with acute critical illness (for an overview see Arafah, 2006). Nevertheless, recently it has been argued that critically ill patients, especially those with sepsis, may have relative adrenal insufficiency despite cortisol values within or above the given reference range (for an overview see Thomas and Fraser, 2007).

While sepsis-induced relative hypocortisolism may be largely attributable to primary adrenal insufficiency, there is concern that traumatic brain injury (TBI) and aneurysmal subarachnoid hemorrhage (SAH) may also compromise the HPAA on the hypothalamo–pituitary level, with the potential consequence of secondary adrenal insufficiency (AI) (Cooper and Stewart, 2007). Within the last years, SAH has been identified as a major and formerly underestimated cause of hypopituitarism as a long-term complication of the bleeding. Adrenocorticotrophic hormone (ACTH) insufficiency and other neuroendocrine disturbances have been confirmed in 47.5% of 102 SAH survivors investigated between 5 months and several years after the acute event (for an overview see Schneider et al., 2007), and very recently, in a retrospective study, AI was identified in a high proportion of SAH patients who were nonresponsive to vasopressor therapy and received cosyntropin for the evaluation of AI within 14 days of the hemorrhage (Weant et al., 2008). Another study on the HPAA in the acute phase of aneurysmal SAH identified elevated total and free serum cortisol values on the first day after SAH as compared to a control group of patients undergoing elective surgery for an unruptured aneurysm (Bendel et al., 2008), but argued since SAH severity did not affect cortisol concentrations, this may indicate relative adrenal insufficiency in those patients with a more severe bleeding.

However, in patients with severe brain injury such as SAH, the pattern of cortisol secretion and regulation is still poorly understood. One study, addressing cortisol dynamics by collecting a morning and an evening cortisol and cortisol binding globulin (CBG) values in 15 patients after TBI (n = 9) and SAH (n = 6) showed that the total morning cortisol in the majority of these patients was similar to the reference range in normal individuals, but that CBG was significantly decreased, leading to an increased free cortisol level (Savaridas et al., 2004).

It was the aim of the present study to broaden and augment the knowledge of cortisol dynamics in the acute phase of aneurysmal SAH as a building block for establishing a tool for the assessment of AI in these patients. Therefore, we investigated the circadian rhythm of cortisol and CBG, as well as ACTH, and set parameters of cortisol release in relation to clinical events and complications. As interleukin-6 (IL-6) constitutes a potentially important factor of extra-ACTH cortisol release, we also assessed this parameter (Bornstein and Chrousos, 1999).

Methods

Patients

All patients admitted for aneurysmal SAH between October 2005 and August 2006 to the Department of Neurosurgery, RWTH Aachen University Hospital, Aachen were screened for eligibility for this study. Exclusion criteria were pregnancy, glucocorticoid medication on admission to hospital or during treatment, prior pituitary insufficiency, as well as being aged below 18 or over 70 years. A total of 22 patients could be included in this prospective observational study (14 females, 8 males; mean age 47.2 years, range 25–69 years). The study was approved by the local ethics committee of the University of Technology, RWTH Aachen, and was carried out according to the Declaration of Helsinki. Written informed consent was given in all cases. In patients too ill to give consent at the time of ICU admission, consent by proxy was obtained.

No patient received dopamine, ketoconazole, or etomidate. A total of 15 patients received norepinephrine for blood-pressure maintenance or prophylactic elevation of blood pressure in case of dopplersonographically proven elevated blood-flow velocities of intracranial vessels (induced hypertension, see Kreitschmann-Andermahr and Gilsbach, 1996). Antibiotic drugs for the treatment of bacterial infection were administered to 15 patients during their stay on the ICU.

Table 1 shows the patient characteristics with sex, age, aneurysm location, clinical status on admission to hospital according to the grading system of Hunt and Hess (HH) (Hunt and Hess, 1968), the amount of subarachnoid blood visualized by computed tomography (CT) according to the Fisher scale (Fisher et al., 1980), and the outcome at the time of hospital discharge as measured by the Glasgow Outcome Scale (GOS) (Jennett and Bond, 1975). Seven days (t ₁) and at least 14 days (t ₂) after the bleeding, the GCS (Teasdale and Jennett, 1974) was recorded for each patient.

Table 1.

Patient Characteristics

							CFSC rhythm
Pat. no	Sex	Age	HH grade	Fisher CT score	GOS	Aneurysm location	t₁	t₂
1	F	44.9	4	3	3	MCA	Missing	Missing
2	F	46.7	5	3	5	ACoA	Regular	Regular
3	F	48.7	4	3	3	ACoA	Flat	Abnormal
4	F	47.1	4	3	4	PICA	Abnormal	Regular
5	F	69.1	2	2	4	MCA	Flat	Abnormal
6	M	49.1	4	3	3	ACoA	Abnormal	Abnormal
7	F	51.6	3	2	4	MCA	Abnormal	Abnormal
8	M	52.0	1	2	5	ACoA	Regular	Regular
9	F	39.9	5	3	5	A1	Abnormal	Regular
10	M	49.8	4	3	2	MCA	Abnormal	Abnormal
11	F	33.4	5	3	5	PCoA	Regular	Regular
12	M	31.7	1	2	5	ACoA	Regular	Regular
13	F	53.5	1	2	5	AChorA	Abnormal	Abnormal
14	F	55.6	2	1	4	PICA	Abnormal	Abnormal
15	F	26.8	2	2	5	PICA	Abnormal	Abnormal
16	F	54.0	3	2	4	ACoA	Abnormal	Abnormal
17	M	48.1	2	1	5	ACoA	Regular	Abnormal
18	M	59.1	5	3	4	ACoA	Abnormal	Abnormal
19	F	43.7	4	3	5	MCA	Flat	Missing
20	M	25.1	3	3	5	ACoA	Regular	Regular
21	F	69.5	3	3	5	BAS	Flat	Regular
22	M	38.2	1	2	5	ACoA	Flat	Regular

HH grade: 1 = asymptomatic or minimal headache and slight nuchal rigidity; 2 = moderate to severe headache, nuchal rigidity, and no neurological deficit other than cranial nerve palsy; 3 = drowsiness, confusion, or mild focal deficit; 4 = stupor, moderate to severe hemiparesis, possibly early decerebrate rigidity, and vegetative disturbances; and 5 = deep coma, decerebrate rigidity, moribund appearance. Fisher CT score: 1 = no blood detected; 2 = a diffuse disposition or thin layer with all vertical layers of blood (interhemispheric fissure, insular cistern, ambient cistern) less than 1 mm thick; 3 = localized clots and/or vertical layers of blood 1 mm or greater in thickness; and 4 = diffuse or no subarachnoid blood, but with intracerebral or intraventricular blood. GOS grade: 1 = death; 2 = persistent vegetative state; 3 = severe disability (patient depends upon others for daily support due to mental or physical disability or both); 4 = moderate disability (disabled but independent); and 5 = good recovery. Aneurysm location: MCA = middle cerebral artery; ACoA = anterior communicating artery; PICA = posterior inferior cerebellar artery; VA = vertebral artery; A1 anterior cerebral artery; PCoA = posterior communicating artery; AChorA = anterior choroidal artery; ICA = internal carotid artery; BA = basilar artery.

The symptomatic aneurysm could be identified in all patients by angiography, and was treated in all patients. In addition to the symptomatic aneurysm, in six patients multiple aneurysms were present; none of those aneurysms were treated within the study period. A total of 15 patients underwent microsurgery and seven patients endovascular therapy with coil embolisation. The choice of occlusion method was interdisciplinarily decided upon and based on aneurysm characteristics and the patient's preference. In all cases, the ruptured aneurysm was occluded within 48 h after the initial SAH.

The duration of ICU stay in this cohort was 6–37 days (mean 19 days). All patients received intensive-care management by experienced neurointensivists. Nimodipine was administered intravenously for 14 days (2 mg/h) and a daily transcranial ultrasound (TCD) for the detection of cerebral vasospasm was performed. Cerebral vasospasm was defined as a TCD mean cerebral blood flow in the middle cerebral and/or the internal carotid artery >120 cm/s with a Lindegaard index >3 (Lindegaard et al., 1988); vasospasm was diagnosed in 11 patients. At t ₁, six patients were analgosedated. Of them, five were fully ventilated, and one additional patient received assisted ventilation but was not analgosedated. At t ₂, three patients were analgosedated. Of them, one was fully ventilated, and two received assisted ventilation. Above this, two patients were ventilated without analgosedation (one patient assisted ventilation, one patient full ventilation). Patients who received assisted ventilation are classified as “ventilated” in the following calculations. The neurosurgical ICU was fitted with large daylight windows. Additional lighting was dimmed overnight, and disturbances were kept to a minimum. In all patients who did not require parenteral nutrition, regular mealtimes were observed. This meant that external time cues existed, which helped patients to keep in synchrony with normal circadian rhythms.

Methods

In all patients cortisol, CBG, ACTH, and IL-6 were assessed at three different time points: (t ₀) at admission to the hospital; at 8:00 a.m. on the seventh day after the bleeding (t ₁, mean 7.5 days, SD 0.9), and at 8:00 a.m. at least 14 days (mean 17.1, SD 4.7) after the bleeding (t ₂).

All blood samples were collected via an existing indwelling central or peripheral line. Furthermore, at t ₁ and t ₂, diurnal cortisol profiles were assessed.

For the assessment of these diurnal cortisol profiles, serum cortisol and CBG measurements were performed at t ₁ and t ₂ at four time points each (8, 12, 16, 20 h). To calculate the amount of free cortisol in plasma (calculated free serum cortisol, CFSC), we used Coolens' equation (Coolens et al., 1987).

In order to gain an overview of the course of cortisol secretion, the circadian variation of CFSC was rated as a normal or abnormal: normal diurnal cycles included regular cycles with clearly raised morning values and a consistent decline over the day, resulting in the lowest value in the evening (“regular” rhythm). CFSC profiles were also labeled as normal when they exhibited only a minimal diurnal variation (flat cycles) (Stone et al., 2001). Flat cycles presumably occur in at least 10% of the healthy population, and connections to health status are not evident (Stone et al., 2001). Cortisol slopes that did not follow a normal pattern, that is, with a reversed slope, or profiles that followed a strong zigzag pattern were classified as abnormal. Rating was performed only if there were at least three measurements throughout the day. Allocation to the different profiles was done by visual inspection by an investigator blinded to the clinical variables of the patients. All curves could clearly be allocated to one of the three types (regular, flat, abnormal) of secretion pattern. In addition to visual inspection of the slope, the area under the curve (AUC, here calculated in respect to ground) (Pruessner et al., 2003) was calculated for CFSC.

In order to classify the course of the illness, a complication index that ranged from 0 to 2 was created. This clustering of complications was performed due to the small sample size and in order to avoid multiple comparisons, which would have led to an increased type-1 error. An uneventful clinical course without any relevant medical complications was rated as 0 (n = 10). Mild to moderate complications such as pneumonia not requiring reintubation and which could easily be treated with antibiotics were rated as 1 (n = 7), while severe complications such as rebleeding, sepsis, or severe pneumonia requiring reintubation were rated as 2 (n = 5). The complication index was rated independently by two of the authors (IKA and MR).

Analytes and assays

All analyses were carried out in the central laboratory of the University Hospital Aachen. Reference values for the analytes in the normal population as provided by the laboratory are listed in parentheses behind the respective assays.

Specific quantitative detection of cortisol in serum was carried out using an automated immunometric chemiluminescence assay (ADVIA Centaur^®; Bayer Health Care Diagnostics, Fernwald, Germany; reference value: 171–536 nmol/L). Serum CBG concentrations were determined by quantitative ¹²⁵I-CBG radioimmunoassay (RIA) (ZenTech, Angleur, Belgium; reference value men: 33.3–61.3 mg/L; reference value women: 21.8–82.6 mg/L). Serum IL-6 was measured by chemiluminescent sequential immunometric assay (Immulite 1000; DPC Biermann, Bad Nauheim, Germany; reference value: <15 ng/L). Due to technical reasons, the method of ACTH determination had to be changed twice during the study period. Until January 2006, the assay from Nichols Institute Diagnostics, San Clemente, CA (reference value: 2.0–11.5 pmol/L) was used; from January to May 2006, the assay from Diasorin, Saluggia (Vercelli), Italy (reference value: 0.9–6.1 pmol/L) was used. Since May 2006, ACTH was determined via E170 (Roche GmbH, Basel, Switzerland; reference value: 1.5–14.7 pmol/L).

Statistical analysis

SPSS statistical software (V17.0; SPSS, Inc., Chicago, IL) was used for statistical analysis. Due to the sample size (n = 22), nonparametric statistics were calculated. To control for possible confounding effects of norepinephrine administration, Mann–Whitney U-tests with the dependent variables CFSC (08:00, 12:00, 16:00, 20:00), AUC, and CFSC morning/evening ratio were calculated. U-tests were also calculated in order to examine possible group differences between normal vs. abnormal diurnal cortisol profiles in respect to clinical parameters, and to compare patients of both treatment modalities (coiling/clipping). To test whether ventilation and sedation have an influence on the pattern of diurnal cortisol secretion, a chi-square test (Monte Carlo method used due to small cell sizes) was performed with the variables “sedated and ventilated: yes/no” and “normal rhythm: yes/no.” In order to analyze changes over time in the total amount of daily secreted cortisol (AUC), Wilcoxon signed-rank tests were calculated. Kruskal–Wallis H-tests were calculated in order to explore possible group differences between the different ACTH gradings (below, in, and above reference range) and 08:00 CFSC values at t ₁ and t ₂. This analysis for the exploration of the relationship between cortisol and ACTH release was performed owing to the change of ACTH assays during the study period, which rendered the calculation of correlation coefficients impossible. Spearman rank–order correlations were calculated to analyze relationships between single CFSC levels and ordinally scaled clinical variables, including IL-6 as a possible marker for ACTH independent cortisol secretion. In all tests, p values of <0.05 were considered statistically significant.

Results

Basal serum cortisol concentration at t ₀ (admission) ranged between 189 and 857 (mean 536.05), at t ₁ (08:00) between 261 and 1189 (mean 570.36), and at t ₂ (08:00) between 151 and 1123 nmol/L (mean 430.50; Table 2). The mean cortisol values at t ₂ were significantly lower than the 08:00 values at t ₁ (Wilcoxon signed-rank test p = 0.013).

Table 2.

Concentrations of Cortisol and IL-6

	t₀	t₁	t₂
Cortisol nmol/L (08:00)
Mean (SD)	536.05 (203.48)	570.36 (225.61)	430.50 (217.21)
Range	189–857	261–1189	151–1123
Below/within/above reference range	0/12/9	0/9/13	1/17/4
CFSC ng/mL (08:00)
Mean (SD)	–	16.01 (9.84)	10.62 (7.82)
Range	–	4.2–41.8	1.7–39.6
CBG mg/L (08:00)
Mean (SD)	–	44.90 (5.84)	43.90 (8.98)
Range	–	37.0–58.6	25.0–56.6
Below/within/above reference range	–	0/21/0	1/20/0
IL-6 ng/L (08:00)
Mean (SD)	32.24 (49.21)	27.71 (44.98)	15.22 (19.68)
Range	3–180	3–210	2–82
Within/above reference range	10/11	14/8	16/6

t ₀, hospital admission; t ₁, 7 days after the hemorrhage; t ₂, 14–21 days after the hemorrhage. Reference values: cortisol – 171–536 nmol/L; CBG women – 21.8–82.6; men – 33.3–61.3 mg/L; IL-6 – 0–15 ng/L.

Cortisol values were above the reference range (08:00 samples) provided by the laboratory for healthy individuals at t ₁ in 13 patients and at t ₂ in four patients. The concentration was not below the normal reference range at t ₁ in any patients and at t ₂ in one patient (Table 2). CBG concentrations were slightly below the reference range in one patient at t ₁ (16:00 measurement: 28.4 mg/L; reference range: 33.3–61.3) and another patient at t ₂ (08:00 measurement: 30.5 mg/L; reference range: 33.3–61.3). All other measures at t ₁ and t ₂ were within the reference range. CBG was measured together with cortisol four times a day. For better clarity and comparability of the data, statistics in Table 2 are given for 08:00 values only. Additionally, CFSC at 08:00 is given in Table 2. ACTH concentration was above the respective reference range for healthy individuals at t ₀ in one patient, at t ₁ in four patients, and at t ₂ in two patients. ACTH was measured below the reference range at t ₀ in six patients, at t ₁ in six patients, and at t ₂ in two patients. Kruskal–Wallis H-tests showed an association between 08:00 ACTH classifications (above, within, or below reference range) in respect to 08:00 CFSC values at t ₁ (p = 0.038), but not at t ₂ (p = 0.091). This means that high CFSC values in SAH patients were not consistently associated with high ACTH values. Correlation analysis between IL-6 and CFSC levels was not statistically significant at any time point. The correlation of CFSC and serum cortisol for all measurements was statistically significant (Spearman correlation, r = 0.954, p < 0.0001).

Diurnal cortisol rhythm

Serum cortisol profiles could be obtained in 21 patients at t ₁, and 20 patients at t ₂, and were each allocated to the normal or abnormal different secretion patterns as described above (see Table 3). Flat profiles, which were allocated to the normal rhythms exhibited a standard deviation (SD) of the respective CFSC patterns of <2 ng/mL (in all cycles, intrasubject SD ranged between 0.66 and 25.13 ng/mL). In one patient, both diurnal profiles were missing due to technical reasons. Of the 11 patients with a normal profile at t ₁ (six regular, five flat), three switched to abnormal patterns at t ₂. Seven patients still had normal patterns at t ₂, one was missing. Of the 10 patients with an abnormal profile at t ₁, eight still had an abnormal profile at t ₂, and two patients now exhibited a normal (regular) CFSC profile (Table 1). At t ₁ and t ₂, patients that were sedated and ventilated exhibited abnormal diurnal profiles significantly more often than patients able to breathe on their own (p = 0.012 and p = 0.001 respectively).

Table 3.

Allocation to the Different Patterns of Diurnal Cortisol Secretion

		t₁ n (%)	t₂ n (%)
Normal	Regular	6 (27.3%)	9 (40.9%)
	Flat	5 (22.7%)	0 (0.0%)
Abnormal	Reversed/undefined	10 (45.5%)	11 (50.0%)
	Missing	1 (4.5%)	2 (9.1%)

t ₁, 7 days after the hemorrhage; t ₂, 14–21 days after the hemorrhage.

The total amount of daily CFSC as assessed by AUC ranged between 49.08 and 339.49 at t ₁ (mean: 135.89), and between 31.55 and 191.62 at t ₂ (mean: 95.03). In contrast to the 08:00 cortisol value, the mean AUC was not significantly higher at t ₁ than at t ₂ (Wilcoxon signed-rank test p = 0.062).

Mann–Whitney U-tests showed that patients with a normal diurnal profile of CFSC at t ₁ had a shorter ICU stay (p = 0.038), a lower complication index (p = 0.043), a more favorable GCS at t ₁ (p = 0.043), and a more favorable outcome (GOS, p = 0.051) than patients with an abnormal diurnal profile. Neither HH score (p = 0.468) nor Fisher CT score (p = 0.863) differed between both groups. At t ₂, the appropriate GCS (p = 0.016) as well as GOS (p = 0.012) differed significantly between both groups. Group differences did not reach statistical significance for ICU stay (p = 0.182), complication index (p = 0.112), HH score (p = 0.766), and Fisher CT score (p = 0.175).

CFSC values at 08:00 and at 20:00 alone were not associated with ICU days, complication index, GOS, Fisher CT, or HH score at t ₁. Only morning CFSC and GCS exhibited a small correlation (r = 0.439, p = 0.046, Spearman correlations). At t ₂, however, evening values correlated with the number of days spent on ICU (r = 0.646, p = 0.005), complication index (r = 0.544, p = 0.020), GOS (r = −0.668, p = 0.002), and GCS at that day (r = −0.559, p = 0.016, all Spearman correlations), indicating a connection between elevated evening values and a longer ICU stay, a higher complication index, a less favorable outcome, and a worse level of consciousness at that day. The Fisher CT score and HH score were not associated with morning or evening values of CFSC at t ₂. The total amount of CFSC, defined as AUC, was not associated with any of the above-mentioned variables at both time points.

Noteworthily, a separate analysis of vasospasm in relation to AUC revealed a reduced amount of CFSF in the group of patients with vasospasm at t ₂ (p = 0.019, U-test) but no group differences in respect to other CFSC-related parameters such as evening value, morning value, or secretion pattern. There was no significant difference in cortisol values between patients with and without norepinephrine medication (U-test, n.s., data not shown). Treatment condition was neither significantly associated with the dependent variables vasospasm, GCS, complication index, or outcome (GOS), nor with different markers of cortisol secretion (AUC, 08:00 CFSC, 20:00 CFSC, normal diurnal rhythm vs. abnormal rhythm) at t ₁ or t ₂ (U-test, n.s., data not shown).

Discussion

This investigation was performed to examine cortisol dynamics prospectively in patients with aneurysmal SAH in the acute phase of the illness. The analysis of the basal hormones post-admission and at 7 days and ≥14 days revealed the following insights. Cortisol values at all three time points showed a wide variability. At t ₁, 08:00 serum cortisol was significantly higher than at t ₂. At t ₁, this parameter was, however, not associated with clinical parameters of disease severity or outcome, resembling the findings of Bendel and associates (2008) who assessed serum free and total cortisol concentrations in 30 patients shortly after SAH and found them unaffected by the severity of SAH. Furthermore, 08:00 CFSC and total CFSC output (AUC) exhibited a significant correlation at t ₁ but not at t ₂. This weak association may be due to the distorted diurnal variation of CFSC in some patients, leading to the fact that the 08:00 value is not the highest value throughout the day, or to the oftentimes strong variation of CFSC throughout the day including high peaks, possibly leading to imprecise estimations of the total amount of cortisol released. In the latter case, variables less sensitive to abrupt variation may be more useful, such as measurements of cortisol in 24-h urine. A reason for the missing association between AUC and clinical variables may be not that the total amount of cortisol is important for the stress response in the acute phase after aneurysmal SAH, but a normal pattern of cortisol secretion. In our study, a normal diurnal profile of CFSC at t ₁ was connected to a less severe illness and a more favorable outcome, and at t ₂ to a more favorable outcome only. This may be due to abnormally raised evening values; 2–3 weeks after the bleeding, elevated CFSC values in the evening were connected to a more severe illness and a poorer outcome.

There was, moreover, no consistent relationship between ACTH values in respect to reference ranges and 08:00 CFSC levels or a significant association between IL-6 and CSFC. In the traditional view of HPA regulatory mechanisms, an adaptation to stress would be suggested by parallel increases in ACTH and cortisol levels according to the degree of stress (Groeneveld et al., 2009). In our SAH cohort, this was only observed at t ₁, although this result must be viewed with caution due to the limitation of the twice-changed ACTH assay. All the same, Bendel and associates (2008) also reported a dissociation of ACTH and cortisol levels at most time points in the author's cohort. However, it is important to keep in mind that a separation of ACTH and cortisol secretion has clinical relevance. This concept must be integrated into the appraisal of the HPA axis during acute illness, since 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 (for an overview see Bornstein et al., 2008). However, in this study, there was also no association observed between IL-6 as a possible marker for extra-ACTH cortisol secretion and CFSC levels.

It is conceivable that cortisol secretion in patients after aneurysmal SAH is influenced by central dysregulation that may be caused by decreased hypothalamo–pituitary perfusion due to vasospasm, irritation of the pituitary stalk due to subarachnoid blood or hypothalamo–pituitary hemorrhage. Indeed, one neuropathological investigation has confirmed a high incidence of hypothalamic lesions in patients who died shortly after aneurysmal SAH (Crompton, 1963). Of note, in our investigation, we observed that patients with cerebral vasospasm as assessed by TCD – which may indicate hypothalamic hypoperfusion – exhibited a significantly lower amount of total CFSC (AUC) 14 to 21 days after the hemorrhage. Additionally, since SAH constitutes a trauma to the brain, it can be assumed that the disturbance of cortisol profiles reflects an alteration of central cortisol secretion and not an alteration of cortisol clearance.

Diurnal cortisol output in the acute phase may be influenced by sleeping patterns (Lasikiewicz et al., 2008). The assessment of sleep–wake cycles in patients on an ICU requires close monitoring, and is usually not performed. Therefore, in our study, no data concerning sleep patterns were available in the investigated patients. However, as pointed out in the methods section, external zeitgebers were available in the ICU setting to allow patients to keep up circadian rhythmicity. Patients that were sedated and ventilated exhibited abnormal diurnal cortisol profiles more often than patients without this intervention. Whether this is due to the lack of a normal sleeping pattern, or if this is just a marker for disease severity, remains to be clarified.

In our study, the diurnal cortisol profile was abolished in about half of the SAH patients 7 days and also ≥14 days after the acute event. This results in a very important finding for clinical routine, namely that a random cortisol sample drawn sometime during the course of a day cannot be regarded as representative of the course or the appropriateness of daily cortisol secretion in patients shortly after SAH, and therefore is of no avail in evaluating the HPAA response. In our sample, this also holds true for the common routine parameter, the 08:00 cortisol value. We, therefore, cannot approve the use of one single basal or random cortisol sample that is still advocated by some authors as a marker to define relative AI in critically ill patients (Arafah, 2006; Cooper and Stewart, 2007), and which has been used among others to define AI in patients with SAH in a recent study (Weant et al., 2008).

It has been proposed that measured serum total cortisol levels in critically ill patients can be misleadingly lower than anticipated due to decreased concentrations of CBG and albumin, resulting in the incorrect conclusion that adrenal function is impaired (Hamrahian et al., 2004). This led us to investigate CFSC. However, in our sample, CBG was below the reference range in 2 of 154 samples only. Additionally, correlation coefficients between serum cortisol and CFSC were high, a finding that is also reflected in the recently published study by Bendel and associates (2008). This suggests that total serum cortisol probably is a sufficient marker for the evaluation of cortisol status in patients shortly after SAH, but that a singular measurement of CBG may be useful to identify patients with severe CBG deficiency.

Despite the efforts to establish clear rules for the assignment of specific secretion patterns, our method contains a degree of subjectivity. In this cohort, however, all patterns were clearly recognizable; flat cycles had a SD of <2 ng/mL and normal cycles all were dichotomously falling throughout the day. Still, a study on clear rules to differentiate between normal and abnormal diurnal cycles is called for.

Conclusion

In summary, in our study we could not identify an association between morning or total cortisol output and markers of disease severity and outcome in patients shortly after SAH. We could, however, draw an important association between abolished diurnal profile of CFSC and elevated CFSC evening values and measures of disease severity and outcome, leading to the conclusion that other measures of cortisol output besides random cortisol values are needed to make an appraisal of the HPAA in these patients. Due to the small sample size, these findings must be regarded as preliminary. Further prospective investigations of the HPAA in the wake of SAH, which include biochemical and clinical assessments as well as long-term follow-ups, are necessary to identify hormonal risk factors for later adrenal insufficiency or impaired recovery after SAH, and to improve diagnosis and treatment strategies for neuroendocrine dysfunction after SAH.

Footnotes

Acknowledgments

The authors wish to acknowledge the support of the ICU staff for the study. This work was supported in part by a grant from the START program of the medical faculty of the RWTH Aachen.

Author Disclosure Statement

No competing financial interests exist.

References

Arafah

B.M.

2006. Hypothalamic pituitary adrenal function during critical illness: Limitations of current assessment methods. J. Clin. Endocrinol. Metab., 91:3725–3745.

Bendel

, Koivisto

, Ruokonen

, Rinne

, Romppanen

, Vauhkonen

, Kiviniemi

, Uusaro

2008. Pituitary-adrenal function in patients with acute subarachnoid haemorrhage: A prospective cohort study. Crit. Care 12,, R126.

Bornstein

S.R.

, Chrousos

G.P.

1999. Clinical review 104: Adrenocorticotropin (ACTH)- and non-ACTH-mediated regulation of the adrenal cortex: Neural and immune inputs. J. Clin. Endocrinol. Metab., 84:1729–1736.

Bornstein

S.R.

, Engeland

W.C.

, Ehrhart-Bornstein

, Herman

J.P.

2008. Dissociation of ACTH and glucocorticoids. Trends Endocrinol. Metab., 19:175–180.

Coolens

J.L.

, Van

B.H.

, Heyns

1987. Clinical use of unbound plasma cortisol as calculated from total cortisol and corticosteroid-binding globulin. J. Steroid Biochem., 26:197–202.

Cooper

M.S.

, Stewart

P.M.

2007. Adrenal insufficiency in critical illness. J. Intensive Care Med., 22:348–362.

Crompton

M.R.

1963. Hypothalamic lesions following the rupture of cerebral berry aneurysms. Brain, 86:301–314.

Fisher

C.M.

, Kistler

J.P.

, Davis

J.M.

1980. Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning. Neurosurgery, 6:1–9.

Groeneveld

A.J.

, Beishuizen

, Molenaar

2009. Hypothalamo–pituitary–adrenal axis activity after intracranial catastrophies: What is enough? Crit Care 13,, 103.

10.

Hamrahian

A.H.

, Oseni

T.S.

, Arafah

B.M.

2004. Measurements of serum free cortisol in critically ill patients. N. Engl. J. Med., 350:1629–1638.

11.

Hunt

W.E.

, Hess

R.M.

1968. Surgical risk as related to time of intervention in the repair of intracranial aneurysms. J. Neurosurg., 28:14–20.

12.

Jennett

, Bond

1975. Assessment of outcome after severe brain damage. Lancet, 1:480–484.

13.

Kreitschmann-Andermahr

, Gilsbach

J.M.

1996. Prävention und Therapie des Vasospasmus. Zerebraler Vasospasmus – Neue Aspekte zu Pathogenese und Therapie. Böker

D.K.

Biermann Verlag: Zülpich, 74–86.

14.

Lasikiewicz

, Hendrickx

, Talbot

, Dye

2008. Exploration of basal diurnal salivary cortisol profiles in middle-aged adults: Associations with sleep quality and metabolic parameters. Psychoneuroendocrinology, 33:143–151.

15.

Lindegaard

K.F.

, Nornes

, Bakke

S.J.

, Sorteberg

, Nakstad

1988. Cerebral vasospasm after subarachnoid haemorrhage investigated by means of transcranial Doppler ultrasound. Acta Neurochir. Suppl (Wien), 42:81–84.

16.

Pruessner

J.C.

, Kirschbaum

, Meinlschmid

, Hellhammer

D.H.

2003. Two formulas for computation of the area under the curve represent measures of total hormone concentration versus time-dependent change. Psychoneuroendocrinology, 28:916–931.

17.

Savaridas

, Andrews

P.J.

, Harris

2004. Cortisol dynamics following acute severe brain injury. Intensive Care Med., 30:1479–1483.

18.

Schneider

H.J.

, Kreitschmann-Andermahr

, Ghigo

, Stalla

G.K.

, Agha

2007. Hypothalamopituitary dysfunction following traumatic brain injury and aneurysmal subarachnoid hemorrhage: A systematic review. JAMA, 298:1429–1438.

19.

Stone

A.A.

, Schwartz

J.E.

, Smyth

, Kirschbaum

, Cohen

, Hellhammer

D.H.

, Grossman

2001. Individual differences in the diurnal cycle of salivary free cortisol – a replication of flattened cycles for some individuals. Psychoneuroendocrinology, 26:295–306.

20.

Teasdale

, Jennett

1974. Assessment of coma and impaired consciousness. A practical scale. Lancet, 2:81–84.

21.

Thomas

, Fraser

G.L.

2007. An update on the diagnosis of adrenal insufficiency and the use of corticotherapy in critical illness. Ann. Pharmacother., 41:1456–1465.

22.

Weant

K.A.

, Sasaki-Adams

, Dziedzic

, Ewend

2008. Acute relative adrenal insufficiency after aneurysmal subarachnoid hemorrhage. Neurosurgery, 63:645–649.