Neurocritical Care

, Volume 16, Issue 2, pp 219–223

Circadian Variation in Ictus of Aneurysmal Subarachnoid Hemorrhage

Authors

    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Thomas Bleck
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Siddharth Dugar
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Bichun Ouyang
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Yousef Mohammad
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Sayona John
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Pratik Patel
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Vivien Lee
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Shyam Prabhakaran
    • Divison of Cerebrovascular Disease and Neurocritical CareRush University Medical Center
  • Mark Quigg
    • Department of NeurologyUniversity of Virginia
Original Article

DOI: 10.1007/s12028-011-9640-6

Cite this article as:
Temes, R.E., Bleck, T., Dugar, S. et al. Neurocrit Care (2012) 16: 219. doi:10.1007/s12028-011-9640-6
  • 78 Views

Abstract

Background

Temporal patterns in aneurysmal subarachnoid hemorrhage (aSAH) may provide insight into modulation, and therefore, prevention of hemorrhage. We investigated the time of hemorrhage and its relationship to traditional risk factors among patients admitted with aSAH.

Methods

Admitted patients with aSAH were prospectively followed through outcomes and baseline demographics were abstracted through chart review. The group temporal distribution by hour of onset was summarized with cosinor nonlinear least squares. aSAH onset was gathered into night (2300–0500), morning (0500–1100), afternoon (1100–1700), and evening (0500–2300) daily phases. The odds ratio (OR) with 95% CI was calculated for having an aSAH during the morning, afternoon, and evening hours using night as a reference. Multinomial logit models were fitted using aSAH cases across time blocks to determine their associations with different risk factors.

Results

202 patients had the hour of hemorrhage available, and 49 had phase identifiable [total 251: 38 (15%) night, 98 (39%) morning, 58 (23%) afternoon, 57 (23%) evening]. The peak hours of aSAH were between 0700 and 0800 representing 13% of the sample, with a significant cosinor-fitted phase of 7.33(95% CI 5.30, 9.36). For all aSAH cases, morning onset was significantly more common than night onset (OR = 2.58, 95% CI = 1.77–3.75). Nonsmokers were more likely to have aSAH in the morning than smokers (P = 0.043, OR = 3.10, 95% CI = 1.33–7.23).

Conclusions

aSAH occur in a diurnal, morning prevalent pattern regardless of traditional aSAH risk factors. The association of these risk factors with existing onset patterns should be investigated in future studies.

Keywords

Circadian rhythmSubarachnoid hemorrhageRisk factorsNeurocritical CareIctus

Introduction

Aneurysmal subarachnoid hemorrhage (aSAH) is associated with both modifiable and nonmodifiable risk factors. Hypertension, smoking, age, and gender have all been implicated as risk factors for aneurysmal rupture [1]. Cycles of either endogenous or exogenous sources have been described to influence the timing of ischemic [24] and hemorrhagic [58] stroke. For example, a circannual pattern of winter peak and summer nadir occurrences of hemorrhagic stroke have been observed [58]. Weekly or daily patterns have also been observed [27]. Various mediators of diurnal factors that may modulate circadian stroke occurrence may include blood pressure variability [10, 11], physical activity [12], body temperature [13, 14], and hormonal changes [15, 16].

The temporal pattern of stroke occurrence has potential clinical importance because certain risk factors may be modifiable by avoidance of precipitants or manipulation of temporal risk factors. Furthermore, factors associated with the timing of aSAH may also provide clues about the pathogenesis of aSAH.

To describe these potential rhythms in stroke occurrence, we conducted a single-center consecutive case series examination of the temporal patterns of rupture onset among aSAH patients. We also investigated the relationship of temporal patterns with traditional risk factors for aSAH. Frequency patterns were explored to determine if clinical (Hunt Hess Grade) and radiographic (modified Fisher Scale) severity varied by the time of day of rupture.

Methods

Study Population

We studied consecutive patients with aSAH admitted to the Neurosciences Intensive Care Unit (NSICU) at Rush University Medical Center from August 2006 to December 2009 and followed in the subarachnoid hemorrhage outcomes registry. This registry is approved by the hospital institutional review board. aSAH was diagnosed by admission computed tomogram (CT) or by xanthochromia of cerebrospinal fluid if the initial CT scan was nondiagnostic. The presence of an intracranial aneurysm was confirmed through angiography in all cases. Baseline demographics, past medical history, and medical course were abstracted through chart review. Exclusions included secondary SAH from trauma, arteriovenous malformation, or other causes. In patients with “wake up SAH” (symptoms recognized upon awakening), the time of awakening or symptom onset was taken as the time of onset. Patients who did not have within-hour symptom onset time but had time of onset identifiable to a daily phase (morning, afternoon, evening, night) were included as phase identifiable times of onset. Patients who lacked within-hour time of onset, phase identifiable time of onset, or who were “found down” were excluded.

Clinical and Radiological Variables

We recorded baseline demographics, social, past medical history, and symptoms at onset of hemorrhage. All patients received a Hunt–Hess grade (HHG) [17] at time of admission. Admission CT scans were rated using the modified Fisher grading scale (mFS). The modified Fisher SAH CT Scale [18] is an ordinal scale and ranges for 0–IV (0 represents no evidence of SAH blood on CT, I is thin SAH without biventricular blood, II is thin SAH with biventricular blood, III represents thick SAH without biventricular blood, and IV represents thick SAH with biventricular blood).

Statistical Analysis

24 h distributions of aSAH were summarized with cosinor-nonlinear least squares analysis [19]. Other data analyses were performed with commercially available statistical software SAS version 9.1(SAS Institute, Cary, N.C., USA). aSAH onset time was aggregated into the following daily phases: night (2301–0500), morning (0501–1100), afternoon (1101–1700), and evening (1701–2300). The daily distribution of aSAH cases were investigated in each 1 h interval for 24 h. The proportions of aSAH were measured across each phase by age, gender, smoking status, presence/absence of hypertension, HHG, and mFS. Age was dichotomized into <65 years and ≥65 years. HHG was dichotomized into low grade (1–3) and high grade (4–5) aSAH. The odds ratio with 95% confidence interval (CI) of aSAH occurrence in the morning, afternoon, and evening versus the night as a reference was obtained from multinomial logistic regressions. A P value of 0.05 was prospectively taken to indicate significance. These models were also fitted to investigate the relationship between time of rupture and risk factors including hypertension, smoking, age, and gender (expressed as OR).

Results

A total of 202 patients had identifiable onset times within the hour, and additional 49 patients had times identifiable within daily phase, for a total of 251 patients (demographics Table 1). 8 patients had no identifiable time and were excluded from further analysis. Among them, 69% were female, 39% were smokers, 68% had hypertension, and 29% were high grade (HHG 4–5). The mean age of our study group was 56 years. Numbers missing of each variable are listed in Table 1. Reasons for missing data include incomplete charting of baseline demographics and medical history and unavailability of CT imaging for review.
Table 1

Baseline characteristics of study population (n = 251)

Characteristics

 

Gender, n (%)

 

 Male

79 (31)

 Female

172 (69)

Age, mean (sd)

56.2 (13.8)

HHG, n (%)

 

 1

29 (11.6)

 2

91 (36.3)

 3

57 (22.7)

 4

41 (16.3)

 5

32 (12.7)

 Missing

1 (0.4)

Hypertension, n (%)

 

 Yes

171 (68.1)

 No

76 (30.3)

 Missing

4 (1.6)

Smoking, n (%)

 

 Yes

97 (38.6)

 No

149 (59.4)

 Missing

5 (2.0)

Modified Fisher scale, n (%)

 

 1

14 (5.6)

 2

38 (15.1)

 3

127 (50.6)

 4

63 (25.1)

 Missing

9 (3.6)

Of the 202 patients with known hour of aSAH onset, the times of 0700–0800 represented 13% of occurrences (Fig. 1). A cosinor-fitted curve significantly described a daily cycle with acrophase 7.33(95% CI 5.30–9.36). From the total sample, the morning daily phase accounted for 39% of aSAH (Table 2). Morning hemorrhage was higher compared with the night phase (OR = 2.58, 95% CI = 1.77–3.75). Nonsmokers were more likely to have aSAH in the morning than smokers (P = 0.04, OR = 3.10, 95% CI = 1.33–7.23). Gender was of borderline significance with males having a higher odds for rupture (morning vs. night) than females (P = 0.06, OR = 2.62, 95% CI = 0.998–6.87). In all groups, the highest risk of aSAH rupture occurred during the morning phase.
https://static-content.springer.com/image/art%3A10.1007%2Fs12028-011-9640-6/MediaObjects/12028_2011_9640_Fig1_HTML.gif
Fig. 1

Frequency distribution of aSAH onset times: Frequency distribution of aSAH rupture times in 1 h time blocks over 24 h among patients with within-hour identifiable onset times (n = 202). A cosinor-fitted curve demonstrated a daily cycle with an acrophase of 7.33 (7:20 a.m.) with a 95th confidence estimate of 5.30–9.36 (5:18–9:22 a.m.)

Table 2

Proportion of aSAH cases across four time blocks by gender, age, stroke severity (HHG), smoking status, hypertension, and modified Fisher Scale (n = 251)

 

Night (11 p.m.–5 a.m.)

Morning (5–11 a.m.)

Afternoon (11 a.m.–5 p.m.)

Evening (5–11 p.m.)

n (%, 95% CI)

n (%, 95% CI)

n (%, 95% CI)

n (%, 95% CI)

All cases

38(15.1, 10.9–20.2)

98(39.0, 33.0–45.4)

58 (23.1, 18.0–28.8)

57 (22.7, 17.7–28.4)

Gender

 Men

9 (11.4, 5.3–20.5)

36 (45.6, 34.3–57.2)

13 (16.5, 9.1–26.5)

21 (26.6, 17.3–37.7)

 Women

29 (16.9, 11.6–23.3)

62 (36.1, 28.9–43.7)

45 (26.2, 19.8–33.4)

36 (20.9, 15.1–27.8)

Age

 <65

24 (12.9, 8.5–18.6)

75 (40.3, 28.9–43.7)

43 (23.1, 19.8–33.4)

44 (23.7, 15.1–27.8)

 ≥65

14 (21.5, 12.3–33.5)

23 (35.4, 23.9–48.2)

15 (23.1, 13.5–35.2)

13 (20.0, 11.1–31.8)

SAH severity

 HHG = 1–3

27 (15.3, 10.3–21.4)

71 (40.1, 32.8–47.7)

43 (24.3, 18.2–31.3)

36 (20.3, 14.7–27.0)

 HHG = 4–5

11 (15.1, 7.8–25.4)

27 (37.0, 26.0–49.1)

14 (19.2, 10.9–30.1)

21 (28.8, 18.8–40.6)

Smoking

 Yes

21 (21.7, 13.9–31.2)

36 (37.1, 27.5–47.5)

20 (20.6, 13.1–30.0)

20 (20.6, 13.1–30.0)

 No

16 (10.7, 6.3-16.9)

62 (41.6, 33.6–50.0)

35 (23.5, 16.9–31.1)

36 (24.2, 17.5–31.9)

Hypertension

 Yes

30 (17.5, 12.2–24.1)

61 (35.7, 28.5–43.3)

40 (23.4, 17.3–30.5)

40 (23.4, 17.3–30.5)

 No

7 (9.2, 3.8–18.1)

36 (47.4, 35.8–59.2)

16 (21.1, 12.5–31.9)

17 (22.4, 13.6–33.4)

Modified Fisher scale

 1

2 (14.3, 1.8–42.8)

5 (35.7, 12.8–64.9)

3 (21.4, 4.7–50.8)

4 (28.6, 8.4–58.1)

 2

7 (18.4, 7.7–34.3)

14 (36.8, 21.81–54.0)

8 (21.1, 9.6–37.3)

9 (23.7, 11.4–40.2)

 3

16 (12.6, 7.4–19.7)

52 (40.9, 32.3–50.0)

31 (24.4, 17.2–32.8)

28 (22.1, 15.2–30.3)

 4

11 (17.5, 9.1–29.1)

25 (39.7, 27.6–52.8)

11 (17.5, 9.1–29.1)

16 (25.4, 15.3–37.9)

Number missing: 1 SAH severity (afternoon group); 5 Smoking (1 Night, 3 Afternoon, 1 Evening); 4 hypertension (1 Night, 1 Morning, 2 Afternoon); 9 Modified Fisher Score (2 Night, 2 Morning, 5 Afternoon)

Discussion

The most important finding of this study is that aneurysms rupture more frequently during the morning phase of the day than other portions of the day. This study is unique in that we demonstrated that the cyclic preponderance of morning ruptures persisted regardless of traditional risk factors such as hypertension, age, smoking status, and gender as well as severity of hemorrhage (HHG and mFS). The circadian modulation of aSAH occurrence, therefore, appears to be at least as important as other traditional risk factors in both potential strategies in treatment and in pathogenesis of aSAH.

Although circadian influences appear important in modulation of aSAH, some limitations remain. First, because patients had their events in their normal environments and not under constant environmental conditions, we have not demonstrated circadian control (sensa stricta). This study, as well as its predecessors [29], does not directly measure whether the observed daily distributions are the result of exogenous or endogenous factors. Possible entraining influences that are external to the system, such as light–dark exposure, mealtimes, medication administration, or other habitual, daily activities could serve to constrain aSAH expression in a diurnal, exogenously mediated pattern. On the other hand, endogenous rhythms, such as circadian rhythms originating in the suprachiasmatic nucleus [20] or the circadian fluctuations of hormones may influence endogenous, circadian timing of aSAH.

We did find an inverse relationship between severity of SAH (HHG) and size of hemorrhage (mFS) which was greatest during the morning hours. During this time period, 79 patients experienced low grade SAH (HHG 1–3), while 77 patients had high fisher scores (mFS 3–4). It is possible that important hormonal systems may modify the relationship between the clinical grade of SAH and size of hemorrhage. One such system that may be disrupted in acute SAH is the hypothalamic–pituitary–adrenal axis (HPA). Some studies have demonstrated that ACTH and stimulus induced cortisol levels rise acutely following aSAH [21], while others have shown that both are elevated following hemorrhage and are not correlated with severity of injury [22]. It is possible that hormonal systems such as the HPA may play a role in the clinical presentation following acute aSAH and the initial response to injury. While disruption of the HPA has been associated with impaired cognitive outcomes following aSAH in some studies [23], a clear relationship has yet to be seen. Future studies should investigate how hormonal systems such as the HPA may influence patterns of acute brain injury following aSAH.

Any study that uses behavior as a means of identifying the time of aneurysm rupture could be biased by favoring recording of events during wakefulness; asymptomatic or mild rupture may be preferentially noted during active periods of the day and may be missed at night. However, the finding that timing of aSAH did not vary by severity (HHG or mFS) mitigates against this possibility. We did not have information pertaining to risk factor duration or severity status in our sample. It is possible that chronicity of diseases such as hypertension or quantitative measures of smoking yield important relationships to aSAH that were not discernable in this study.

A further limitation is that it is possible in those patients with “wake up SAH,” the true time of hemorrhage occurred during the middle of the night rather than when they awoke with symptoms. In addition, we do not have information on those patients who died prior to reaching our institution. As a consequence, a severity and survival bias may have been introduced.

Despite limitations, the present study describes a clear morning propensity for aneurysm rupture that confirms previous observations of not only aSAH, but also hemorrhagic and ischemic stroke. The observation that different stroke mechanisms may share a common circadian pattern suggests that there may also be shared pathophysiologies. First, the findings implicate several hormonal systems that are regulated by the circadian timing system that are also responsive to or facilitate activity. It is possible that hormonal factors related to the hypothalamic–pituitary axis may play a role [4]. Both cerebral aneurysms and systemic aneurysms have been associated with certain endocrinopathies such as panhypopituitarism [21] and hyperprolactinemia [22]. Estrogen has been implicated in aneurysm enlargement [16]. Second, changes in blood pressure have been found to be correlated with the circadian pattern of hemorrhagic stroke [10, 11] with increases in blood pressure being highest in the morning and afternoon [11]. Interestingly, we found that both nonhypertensives and hypertensives had higher proportions of morning rupture times. This may reflect that circadian variation in blood pressure and not hypertension has more of an important role in aneurysm rupture times as reported previously [7]. It has been found that hypertensive patients who also smoked had larger morning blood pressure surges than nonsmoking hypertensives [23]. Although we found that nonsmokers had a larger proportion of morning ruptures than smokers, both groups exhibited the same circadian pattern as other traditional risk factors. It may be that endogenous or exogenous factors pertaining to this time period have a stronger influence among nonsmokers. Third, physical activity may play an important role in the timing of aSAH. Certain stressors and physical activities experienced during the morning hours may play an important role in aneurysm rupture. The sudden transition from a state of sleep to one of physical activity may trigger the onset of hemorrhagic stroke [7].

In conclusion, we found that there was a significantly higher proportion of aSAH in the morning hours in our cohort. This relationship was present regardless of gender, age, clinical or radiographic grade and risk factor status. Future studies should investigate how other endogenous or exogenous factors may precipitate aSAH. An understanding in such relationships may serve as a foundation for treatment and counseling of patients with unruptured brain aneurysms.

Acknowledgments

Everyone who made meaningful contributions is listed as author on this manuscript. Bichun Ouyang performed the multinomial logit regression analysis. Mark Quigg performed the cosinor nonlinear least squares analysis.

Copyright information

© Springer Science+Business Media, LLC 2011