Effect of stress‐induced hyperglycemia after non‐traumatic non‐aneurysmal subarachnoid hemorrhage on clinical complications and functional outcomes

Abstract Background Despite having an overall benign course, non‐traumatic non‐aneurysmal subarachnoid hemorrhage (naSAH) is still accompanied by a risk of clinical complications and poor outcomes. Risk factors and mechanisms of complications and poor outcomes after naSAH remain unknown. Our aim was to explore the effect of stress‐induced hyperglycemia (SIH) on complication rates and functional outcomes in naSAH patients. Methods We retrospectively reviewed patients with naSAH admitted to our institution between 2013 and 2018. SIH was identified according to previous criterion. Symptomatic vasospasm, delayed cerebral infarction, and hydrocephalus were identified as main complications. Outcomes were reviewed using a modified Rankin Scale (mRS) at discharge, 3 months, and 12 months. A statistical analysis was conducted to reveal the associations of SIH with complications and outcomes. Results A total of 244 naSAH patients were included in the cohort with 74 (30.3%) SIH. After adjusting for age, gender, hypertension, Hunt and Hess (HH) grade, modified Fisher Scale (mFS), intraventricular hemorrhage (IVH), and subarachnoid blood distribution, SIH was significantly associated with symptomatic vasospasm (p < 0.001, 12.176 [4.904–30.231]), delayed cerebral infarction (p < 0.001, 12.434 [3.850–40.161]), hydrocephalus (p = 0.008, 5.771 [1.570–21.222]), and poor outcome at 12 months (p = 0.006, 5.506 [1.632–18.581]), whereas the correlation between SIH and poor outcome at discharge (p = 0.064, 2.409 [0.951–6.100]) or 3 months (p = 0.110, 2.029 [0.852–4.833]) was not significant. Incorporation of SIH increased the area under curve (AUC) of ROC in the combined model for predicting symptomatic vasospasm (p = 0.002), delayed cerebral infarction (p = 0.024), hydrocephalus (p = 0.037), and 12‐month poor outcome (p = 0.087). Conclusions SIH is a significant and independent risk factor for symptomatic vasospasm, delayed cerebral infarction, hydrocephalus, and long‐term poor outcome in naSAH patients. Identifying SIH early after naSAH is important for decision‐making and treatment planning.


| INTRODUC TI ON
In approximately 15% of spontaneous subarachnoid hemorrhage (SAH) patients, the source of intracranial hemorrhage could not be determined. [1][2][3] These are termed non-traumatic non-aneurysmal SAH (naSAH). 4 Compared with aneurysmal SAH (aSAH), naSAH has an overall benign course of disease. 5 However, some patients with naSAH still develop clinical complications or achieve poor functional outcomes despite their mild condition at admission. [5][6][7] The risk factors and pathophysiological mechanisms of clinical complications and poor outcomes after naSAH remain unknown.

Stress-induced hyperglycemia (SIH) is a transient hyperglycemia
caused by an acute illness. 8 It is an adaptive immune-neurohormonal response to stress, and is often associated with increased morbidity and mortality. 8 Post-SAH hyperglycemia may cause secondary brain damage and cerebral vasospasm. 9,10 One previous study showed that post-aSAH hyperglycemia was associated with the development of symptomatic vasospasm. 11 Juvela et al. reported that hyperglycemia following aSAH was related to delayed cerebral infarction and hydrocephalus. 12 Hyperglycemia after aSAH was also found to be related to an increased risk of poor clinical outcome. 12,13 Additionally, Bian et al. found that a high serum glucose level after aSAH was associated with a high 1-year mortality. 14 However, the prognostic value of SIH in patients with naSAH has not yet been established.
Moreover, these studies did not differentiate between SIH and established diabetes mellitus (DM). Therefore, the objective of this study was to examine the effect of SIH on naSAH patients' complication rates and functional outcomes, and to investigate the prognostic value of SIH for clinical complications and poor outcomes following naSAH.

| Patients and management
We retrospectively reviewed patients suffering from naSAH that were admitted to our institution between January 1, 2013 and December 31, 2018. SAH was diagnosed by computed tomography (CT) or lumbar puncture. Non-traumatic SAH without confirmed bleeding source in cerebral digital subtraction angiography (DSA) examination within 72 h of admission was identified as naSAH. 15 Additionally, patients who met the following criteria were excluded: (1) history of a head injury; (2) history of DM; (3) missing/lost radiological data; and (4) missing/lost laboratory data. All aspects of this study were approved by the institutional board of the Second Affiliated Hospital of Zhejiang University School of Medicine. With their approval, patient consent was not required in this study.
All patients were treated according to SAH guidelines provided by the Neurocritical Care Society and the American Heart Association. 16,17 Nimodipine was used to prevent cerebral vasospasm, and intravenous hydration was received to maintain euvolemia. Hemodynamic values were monitored via electrocardiogram at admission. All patients were not treated with insulin during hospitalization. All patients were treated conservatively rather than invasively.

| Data collection
The baseline characteristics of the patients were reviewed, including age, gender, body mass index (BMI), as well as history of alcohol, smoking, and hypertension. The Hunt and Hess (HH) grade, modified Fisher Scale (mFS), and development of intraventricular hemorrhage (IVH) were used to assess SAH severity. [18][19][20][21] Scores ranging from 3 to 5 for HH grade and 3 to 4 for mFS were considered high.
According to the subarachnoid blood distribution, the patients were stratified into patients with perimesencephalic subarachnoid hemorrhage (PMH) and patients with non-PMH (NPMH). 4,5,15 The characteristics of PMH are as follows: (1) hemorrhage anterior to the midbrain and/or pons; (2) no extension into parenchyma or ventricle deep; (3) incomplete extension into the anterior interhemispheric fissure; (4) possible extension into the basal parts of the sylvian fissures, but not into the lateral sylvian fissures; and (5) no significant IVH. Those who did not meet the bleeding characteristics described above were classified as NPMH. The laboratory data were investigated at admission, including serum glucose, total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), sodium, and potassium. If there was more than one glucose measurement at admission, the median would be taken. Serum glucose levels were monitored during hospitalization. SAH-related complications were reviewed during hospitalization, including symptomatic vasospasm, delayed cerebral infarction, rebleeding, hydrocephalus, and seizure. 22,23 We followed up with all patients in an outpatient clinic or by phone calls. The modified Rankin Scale (mRS) of patients at discharge, 3 months, and 12 months were investigated to assess the functional outcomes. 24 naSAH patients. Identifying SIH early after naSAH is important for decision-making and treatment planning.

K E Y W O R D S
complication, functional outcome, hyperglycemia, neuroendocrine, stress, subarachnoid hemorrhage According to the latest consensus from the American Association of Clinical Endocrinologists and American Diabetes Association, SIH was defined as at least one of the following criteria: (1) an admission serum glucose level of 7.8 mmol/L (140 mg/dl) or more; (2) an inhospital fasting serum glucose level of 7.0 mmol/L (126 mg/dl) or more on 2 or more determinations; and (3) a random serum glucose level of 11.1 mmol/L (200 mg/dl) or more without a prior history of DM. 25

| Outcome measurements
The primary outcomes included the development of clinical complications and poor outcomes. The main clinical complications included symptomatic vasospasm, delayed cerebral infarction, and hydrocephalus. Symptomatic vasospasm referred to either a focal neurological impairment or a decrease of at least 2 points on the Glasgow Coma Scale (GCS) lasting for at least 1 h, which was not immediately apparent after SAH onset, and was not attributable to other causes. 26 Delayed cerebral infarction was diagnosed as a new infarction emerging on CT or magnetic resonance imaging (MRI), which had not originally been present within the first 24-48 h after SAH onset, and was not attributable to other causes. 26 Hydrocephalus was defined as an expansion of the ventricular system on neuroimaging without obstructive cause or typical clinical presentation. 27 Functional outcomes were reviewed at discharge, 3 months, and 12 months using an mRS. Due to the overall favorable prognosis of naSAH patients, mRS scores ranging from 2 to 6 were considered poor outcome. 15 Two senior neurologists independently evaluated all clinical complications and functional outcomes of the patients. If there was a divergence, a third examiner would be used.

| Statistical analysis
Statistical analysis was performed using IBM-SPSS V24.0 (SPSS Inc) with the statistical significance set at p < 0.05. The Shapiro-Wilk test was performed to assess data distribution. Normally distributed variables were expressed as means ± standard deviations (SD).
Categorical variables were expressed as the number of patients

| Patient characteristics
There were 296 patients diagnosed with naSAH in this study.
Thirteen patients had a history of head injury. Twenty-two patients suffered from DM. Ten patients were missing radiological data and seven patients were missing laboratory data. Thus, 244 patients were included in the final cohort, with 74 (30.3%) suffering from SIH ( Figure S1). Among the patients, 108 (44.3%) were women, and the average age was 55.7 ± 11.2 years. Table 1 shows the baseline characteristics, complications, and outcomes of the patients. Patients with SIH had a higher age (p = 0.004) and a higher proportion of hypertension (p = 0.002) and NPMH (p = 0.010). Regarding SAH severity, the HH grade 3-5 (p < 0.001), mFS 3-4 (p = 0.003), and IVH (p = 0.026) all correlated with SIH. In addition, SIH patients were more likely to develop SAHrelated complications, including symptomatic vasospasm, delayed cerebral infarction, and hydrocephalus (all p < 0.001). They also had a higher proportion of mRS 2-6 at discharge (p = 0.013), 3 months (p < 0.001) and 12 months (p < 0.001). Figure S2 shows the subarachnoid blood distribution characteristic, HH grade, mFS, and IVH incidence of SIH and non-SIH patients. The main in-hospital complication rates and mRS distribution of the two groups of patients are shown in Figure S3. The characteristics of PMH and NPMH patients are shown in Table S1.

| Association of variables with clinical complications and functional outcomes
The associations between variables and clinical complications are shown in Table 2. The development of symptomatic vasospasm was significantly associated with NPMH, HH grade 3-5, mFS 3-4, IVH, higher admission serum glucose levels, and SIH (all p < 0.001).
The results of univariate logistic regression analysis for predicting clinical complications and functional outcomes are shown in Table S2 and Table S3.

| Effect of SIH on clinical complications and functional outcomes
SIH was found to be significantly associated with clinical complications and adverse outcomes in our cohort. Therefore, we evaluated the effect of SIH on clinical complications and functional outcomes in multivariate logistic regression analysis. As is depicted in Figure 1,

| Prognostic value of SIH for clinical complications and functional outcomes
The combined models were constructed to predict clinical complications and functional outcomes. The results are displayed in Table 4.

| DISCUSS ION
Although previous studies have explored risk factors for adverse outcomes in naSAH patients, including high clinical and    radiological severity, cerebral edema, and NPMH, the relationship between SIH and outcomes has not been investigated. 28,29 To our knowledge, this is the first study to explore the association of SIH with complication rates and functional outcomes in naSAH.
This study found that SIH after naSAH was significantly and independently associated with the development of symptomatic vasospasm, delayed cerebral infarction, hydrocephalus, and a poor outcome at 12 months after adjusting for demographic data, hypertension history, subarachnoid blood distribution characteristic, and SAH severity. Taking SIH into consideration with risk factors improved the prediction of symptomatic vasospasm, delayed cerebral infarction, hydrocephalus, and 12-month poor outcome after naSAH, although it had limited benefits in the prediction of poor outcomes at discharge or at 3 months. These findings highlight the importance of considering SIH in the decision-making algorithm and treatment planning following naSAH.
SIH is a transient hyperglycemia after acute illness or injury caused by the activation of stress-neuroendocrine axis. 8 Previous research suggested that post-stroke SIH may be a biomarker of stroke severity. 30,31 One meta-analysis involving 16 studies showed that 69% (range, 29 to 100%) of patients suffered a SIH after aSAH. 13 The proportion of SIH in our study cohort (30.3%) was much lower than this proportion, reflecting the low severity of naSAH. In one study, SIH occurred in 32.0% of naSAH patients, which was similar to the proportion in our cohort. 32 An early study confirmed a significant correlation between high serum glucose levels and high clinical severity assessed by HH grade in aSAH patients (p = 0.001). 31 In addition, Santucci et al. reported that SIH after aSAH was significantly associated with radiologically estimated intracranial blood volume (p < 0.001). 33 Our results in naSAH cohort were consistent with these findings. In our study, patients with SIH had higher HH grade (p = 0.001), mFS (p = 0.003) and a higher proportion of IVH (p = 0.026; Figure S2). Additionally, in the characteristic of subarachnoid blood distribution, the proportion of NPMH in SIH patients was higher than that of non-SIH patients (p = 0.010), which may be due to the fact that NPMH was more similar to aSAH with a higher severity. 4,5 These results may indicate the systemic stress response caused by severe brain injury after SAH.
Several studies have explored the relationship between hyperglycemia and complication rates and adverse outcomes after aSAH.
Badjatia et al. found that mean serum glucose levels during hospitalization correlated with the development of symptomatic vasospasm after aSAH (p < 0.001). 11 Juvela and colleagues reported that hyperglycemia following aSAH were related to delayed cerebral infarction and hydrocephalus. 12 A meta-analysis incorporated eight studies for the analysis of the association between hyperglycemia and clinical outcome after aSAH and found that post-aSAH hyperglycemia was associated with an increased risk of poor clinical outcome at 3 or TA B L E 3 Association of variables with functional outcomes at discharge, 3 months, and 12 months 6 months. 13 These results supported our findings. However, these studies only identified patients with hyperglycemia, but did not differentiate between SIH patients and DM patients. Since pre-existing hyperglycemia before SAH onset could not reflect the activation of stress response caused by SAH, our study excluded patients with a history of DM. Moreover, these studies only described the association of hyperglycemia after SAH with functional outcomes within 6 months, but did not explore its relationship with long-term outcomes. Our study identified patients with SIH based on the latest consensus from the American Association of Clinical Endocrinologists and American Diabetes Association, and demonstrated the independent association of SIH with symptomatic vasospasm (p < 0.001), delayed cerebral infarction (p < 0.001), and hydrocephalus (p = 0.008) in the naSAH cohort. After adjusting for demographic data, hypertension history, subarachnoid blood distribution characteristic, and SAH severity, SIH was significantly associated with adverse outcomes at 12 months (p = 0.006), although no significant correlation was found between SIH and poor outcomes at discharge (p = 0.064) or at 3 months (p = 0.110; Figure 1). The poor role of SIH in predicting short-term prognosis may be due to the high proportion of patients with short-term poor outcomes. In our cohort, 81.1% (198/244) of patients had a poor outcome at discharge. However, most (90.4%) of them had an improved outcome at 12 months. The significant association between SIH and long-term poor prognosis may implicate the chronic impairment of SIH in naSAH patients and suggest its value in predicting truly poor outcomes. On the other hand, the association between SIH and short-term poor prognosis may become significant in larger cohorts.
Traditionally, prediction of poor outcomes after aSAH primarily depends on aneurysm size and location and SAH severity assessed F I G U R E 1 Unadjusted and adjusted OR for SIH to evaluate the effect of SIH on clinical complications and functional outcomes. After adjusting for age, gender, hypertension, HH grade, mFS, IVH, and subarachnoid blood distribution characteristic, SIH was still significantly associated with symptomatic vasospasm (A, p < 0.001), delayed cerebral infarction (B, p < 0.001), hydrocephalus (C, p = 0.008), and 12-month poor outcome (F, p = 0.006), but was not significantly associated with discharge poor outcome (D, p = 0.064) and 3-month poor outcome (E, p = 0.110). a Adjusted for age, gender, hypertension, and subarachnoid blood distribution characteristic. b Adjusted for HH grade, mFS, IVH, and subarachnoid blood distribution characteristic. c Adjusted for age, gender, hypertension, HH grade, mFS, IVH, and subarachnoid blood distribution characteristic. HH, Hunt and Hess; IVH, intraventricular hemorrhage; mFS, modified Fisher scale; OR, odds ratio; SIH, stress-induced hyperglycemia by HH grade and mFS. 34,35 However, aneurysm size and location are not applicable in the prediction of poor outcomes in naSAH.
Moreover, naSAH patients usually have minor SAH severity at admission. In our cohort of naSAH, patients with HH grade 1-2 and mFS 0-2 accounted for 87.7% and 77.0%, respectively. Therefore, it is relatively difficult to predict poor outcomes in naSAH patients compared to aSAH patients. Although SIH may be useful in predicting poor outcomes in both aSAH and naSAH patients, 13 the benefits for naSAH patients are obviously greater.
The deleterious effects of activation of stress-neuroendocrine axis after SAH may explain the correlation between SIH and adverse outcomes. The stress response can induce the activation of hypothalamus-pituitary-adrenal axis and sympathetic autonomic nervous system, as well as induce the secretion of glucagon, catecholamines, and corticosteroids. 8 The metabolic disorders caused by these factors may lead to inflammation, systemic damage, and various complications. 8 Sympathetic activation and serum catecholamine elevation after aSAH has been confirmed in previous studies, and were found to be related to symptomatic vasospasm and unfavorable outcomes. 36,37 It was found that inhibition of sympathetic activity by beta-blockers could reduce cerebral vasospasm rates and improve functional outcomes after aSAH. 38,39 In activating protein kinase C after stroke, thereby exacerbating oxidative stress. 49 In another study, hyperglycemia aggravated neuronal apoptosis through the activation of extrinsic caspase cascade via extracellular regulated kinase (ERK) signal pathway after experimental SAH. 9 In addition, hyperglycemia was also found to be related to the activation of platelets and the increase of pro-inflammatory cytokines. 50 Moreover, a recent animal study suggested that hyperglycemia could exacerbate cerebral hemorrhagic transformation by interfering with gut microbiota. 51 In the present study, SIH was a significant and independent risk factor of symptomatic vasospasm, delayed cerebral infarction, hydrocephalus, and 12-month poor outcome in patients with naSAH.  32 Third, the assessments of SAH severity and outcomes may be subjective.
To solve this problem, two senior neurologists performed the evaluations independently. If there was a divergence, a third examiner would be used. Fourth, some patients may have more than one glucose measurement at admission, which may bias the identification of SIH. To address this problem, we would take the median if there was more than one glucose measurement at admission. Fifth, our study did not reveal whether interventions on hyperglycemia after naSAH could assist in reducing risk of the development of complications and poor outcomes. Finally, our study was a single-center retrospective study. Further multi-center prospective studies are needed to verify our findings.

| CON CLUS IONS
This study found that SIH was a significant and independent risk factor for symptomatic vasospasm, delayed cerebral infarction, F I G U R E 2 ROC curve for combined model to predict clinical complications and functional outcomes. In predicting symptomatic vasospasm (A, p = 0.002), delayed cerebral infarction (B, p = 0.024), and hydrocephalus (C, p = 0.037), model 1 had a significantly higher AUC than model 2. In predicting discharge poor outcome (D, p = 0.775) and 3-month poor outcome (E, p = 0.659), the AUC of model 1 was not higher than that of model 2. In the prediction of 12-month poor outcome (F, p = 0.087), the AUC of model 1 was higher than that of model 2, although it was not statistically significant. AUC, area under curve; ROC, receiver operating curve hydrocephalus, and long-term poor outcomes in patients with naSAH. SIH was useful for predicting complications and long-term prognosis of naSAH, although its benefit in the prediction of shortterm prognosis was limited. In addition, this study may allude to the underlying mechanism of stress-neuroendocrine axis in the pathogenesis of naSAH. Our findings highlight the importance of identifying SIH early after naSAH for decision-making and treatment planning.

ACK N OWLED G EM ENT
We are grateful for the data support from the Second Affiliated Hospital of Zhejiang University School of Medicine.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
ZYZ wrote the manuscript; YZ and SC designed the study; AKZ and XYW collected the study data; YJF, SC, and CL revised the manuscript; YBL, HSX, and YJL participated in the design and coordination of the study. All authors read and approved the final version of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.