Endothelial nitric oxide synthase polymorphism (-786T>C) and increased risk of angiographic vasospasm after aneurysmal subarachnoid hemorrhage

Abstract

Background and Purpose

Vasospasm following aneurysmal subarachnoid hemorrhage (SAH) remains a leading cause of death and disability after aneurysm rupture. Decreased availability of nitric oxide (NO) may be crucial in the pathogenesis. We hypothesize that endothelial NO synthase (eNOS) polymorphisms may determine susceptibility to vasospasm in SAH patients.

Methods

We conducted a prospective cohort study of SAH patients, and determined vasospasm by cerebral angiography. We genotyped three eNOS polymorphisms: intron 4 variable-number-tandem-repeat (VNTR), promoter single-nucleotide-polymorphism (-786T>C SNP), and coding SNP in exon 7 (894G>T encoding E298D). Using multivariable logistic regression, we quantified the association of eNOS polymorphisms in patients with vasospasm confirmed by cerebral angiogram.

Results

For the eNOS promoter -786T>C SNP, the presence of the CC genotype compared to any T genotype (CT or TT) was associated with increased odds of vasospasm (OR 2.97, 95%CI 1.32-6.67, p=0.008). No association with vasospasm was observed for the eNOS 894G>T or VNTR polymorphisms.

Conclusions

These findings suggest that genetic variation influencing NO regulation contributes to risk of angiographic vasospasm in patients with SAH. The specific role of the promoter SNP (-786T>C) may determine the effect of NO regulated by this pathway distinct from other known eNOS polymorphisms.

Introduction

Cerebral vasospasm is a common complication after SAH, with a prevalence measured by cerebral angiogram approaching 70% during the first two weeks.¹ Unfortunately, despite maximal medical therapy, 20-30% of patients with angiographic vasospasm will experience delayed ischemic neurologic deficits and cerebral infarcts.² In more recent studies and clinical trials, angiographic vasospasm occurred in 88%, symptomatic vasospasm in 33-66%, and delayed cerebral infarct in 16-44%.^3-7 Among patients who survive with good outcomes, vasospasm contributes to increased hospital length of stay, long-term disability and diminished quality of life.^8-10 Although early detection and initiation of treatment to prevent vasospasm may reduce both acute and long-term disability, current methods do not accurately predict which patients will have clinically significant vasospasm warranting potentially invasive treatments.

The mechanism of injury from vasospasm remains poorly understood, but endothelial nitric oxide (NO) appears important in vasospasm pathogenesis.^11-14 NO is a potent vasodilator and inhibitor of inflammation, smooth muscle proliferation and platelet aggregation.¹⁵ Polymorphisms in endothelial nitric oxide synthase (eNOS) have been associated with increased susceptibility to cerebral aneurysm rupture and risk of vasospasm after SAH in small clinical series.^16-18 In a prospective cohort of SAH patients, we examined the association between angiographic vasospasm and three previously investigated eNOS polymorphisms: a variable-number-tandem-repeat (VNTR) in intron 4,^19-21 a promoter single-nucleotide-polymorphism (−786T>C SNP),^22-26 and a coding SNP in exon 7 (894G>T encoding E298D).^{24, 27-31}

Materials and Methods

Subjects

This was a prospective cohort study of adult patients with aneurysmal SAH followed for acute cardiac events and long-term neurological outcomes. Patients were admitted to a tertiary care referral center with SAH confirmed by non-contrast computed tomography (CT). Lumbar puncture was used in cases with negative head CT but high clinical suspicion of SAH.³² All patients had cerebral aneurysm identified as the source of hemorrhage primarily by cerebral angiogram. Patients were excluded if they had a history of antecedent head trauma or non-aneurysmal source of hemorrhage, and if they could not undergo at least one angiogram during the vasospasm period, typically 3-14 days after SAH. After approval from the institutional review board, consent was obtained from each patient or an appropriate surrogate.

Management of vasospasm

Ruptured aneurysms were treated with either surgical clipping or endovascular coiling within the first 48 hours of admission. All patients were managed in the Neurointensive Care unit. Follow-up head CT and cerebral angiograms were obtained if vasospasm was suspected by elevated TCD velocities and ratios^33-35 or clinical neurologic deterioration not explained by other causes (rebleeding, hydrocephalus, seizures, or metabolic disturbances). All patients received standard medical management for vasospasm with volume resuscitation and induced hypertension.³⁶ Nimodipine was administered to all patients.^{37, 38} Patients who had evidence of angiographic spasm and were failing standard medical therapy were considered for interventional treatment with intraarterial verapamil or transluminal angioplasty.

Measurement of vasospasm and clinical risk factors

For our primary outcome, cerebral vasospasm was defined by the gold standard: cerebral angiography.^{39, 40} Patients were considered to have vasospasm if there was evidence of arterial narrowing compared to the normal vessel caliber (mild <30%, moderate 30-60%, severe 61-99%) as determined by one of three attending neurointerventional radiologists. Follow-up CT scans in a subset of 235 subjects were reviewed by a trained stroke neurologist (NUK, PR). Cerebral infarction on head CT was defined as any new hypodensity in a vascular distribution. Hypodensities that may have been related to post-operative edema or procedural complications were excluded. All outcome measures were performed by investigators blinded to genotype.

We measured known risk factors associated with vasospasm after SAH including age, sex, hypertension, tobacco, and stimulant drug use.^{41, 42} We also included SAH characteristics including clinical severity measured by the presenting Hunt-Hess grade,⁴³ volume of hemorrhage by Fisher group score,⁴⁴ and aneurysm treatment and location.

Polymorphism selection and genotyping

Three eNOS polymorphisms previously reported to be associated with vascular disease were chosen for analysis (Table 1). Genomic DNA was extracted from peripheral blood lymphocytes (Gentra Systems). Polymorphisms were genotyped by template-directed dye-terminator incorporation assay with fluorescence polarization detection (FP-TDI), using the AcycloPrime-FP kit (Perkin-Elmer).⁴⁵ The eNOS VNTR was genotyped by the UCSF Genomics Core on a 3700 DNA Analyzer with GeneScan and GeneMapper v3.0 software (Applied Biosystems). Oligonucleotide primer sequences are shown in Table A (online only). Average genotype success rate was 94% per 96-well plate, and individual plates pass quality control screening above 85%. In 90 subjects, genotype data was available only for the -786T>C SNP, accounting for the variation in number of subjects for the three eNOS genotypes. Genotyping was performed by investigators blinded to vasospasm outcomes.

Table 1. Endothelial nitric oxide synthase (eNOS) polymorphisms.

Polymorphism	Location	Allelesˆ	Clinical rationale	References
−786 T>C	promoter	C, T	C allele associated with: post-SAH vasospasm, coronary artery vasospasm	18-20, 24-28
894 G>T	exon 7	T, G	T allele associated with: coronary artery disease, carotid atherosclerosis, myocardial infarction	26, 29-33
VNTR*	intron 4	a, b	a allele associated with: coronary artery disease, myocardial infarction, cerebral aneurysm rupture, aortic aneurysm rupture	21-23

^*27-base variable-number-tandem-repeat

^ˆC, cytosine; T, thymine; G, guanine; a=4 repeats; b=5 or 6 repeats

Statistical analysis

The primary outcome variable was presence of vasospasm on cerebral angiogram. Our secondary outcome was presence of infarct on delayed CT. Chi-square and Wilcoxon rank-sum tests were used to compare baseline characteristics of subjects with and without angiographic vasospasm and with and without infarct on head CT.

Each polymorphism was tested for adherence to Hardy-Weinberg equilibrium using χ² goodness-of-fit test (P<0.05) for the entire cohort and within each racial subgroup. We examined the relationship between angiographic vasospasm and cerebral infarct with the putative high-risk alleles: “C” allele for −786T>C, the short, 4-repeat (“a”) allele for the VNTR, and the “T” allele for 894G>T (Glu298Asp).

Logistic regression analysis was performed to determine the effect of genotype on angiographic vasospasm or cerebral infarct using odds ratios (OR) and 95% confidence intervals (CI). Clinical covariates with a univariate P value <0.10 were included in multivariable logistic regression models to adjust for potential confounders, in addition to age, gender, and race/ethnicity. Significant interactions among predictor variables and eNOS polymorphisms were evaluated using Mantel-Haenszel χ² test for homogeneity. Additional models were developed to measure the effect of the polymorphisms in whites only, our largest racial subgroup.

To minimize the possibility of type 1 error from multiple comparisons, we used a Bonferroni correction testing three polymorphisms (Bonferroni-adjusted P<0.017). All statistical analyses were performed using commercially-available software (Intercooled STATA® version 9, College Station, TX).

Results

In our cohort of 347 subjects, 71% were female, mean age was 54.7 +/- 13 years, 73% had anterior aneurysm location and 62% were treated with surgical clipping. Mean Hunt-Hess grade was 2.18 and mean Fisher group score was 2.53. Angiographic vasospasm was present in 48%, and cerebral infarct was documented in 35% of 235 subjects with head CT scans obtained after initial presentation. Use of pressors was noted in 53% of subjects. Among those with angiographic vasospasm, subjects were generally younger with anterior aneurysms, and history of stimulant use. Patients were more likely to present with Hunt-Hess grades III or IV and Fisher group 3 (Table 2).

Table 2. Characteristics of SAH patients with and without angiographic vasospasm.

	No Vasospasm n=181	Vasospasm n=166	p value
Mean age in years +/- SD	57±14	52±12	<0.001
Female	125 (69)	121 (73)	0.43
Surgical treatment	102(56)	112(67)	0.19
Anterior location	120(66)	131(79)	0.006
History of hypertension	70(39)	73(44)	0.32
History of smoking	60(33)	55(33)	1.00
History of stimulant use	9(5)	18(11)	0.04
Race/Ethnicity			0.47
Caucasian	119(66)	100(60)
Latino	22(12)	24(14)
Asian	16(9)	23(14)
African-American	19(10)	13(8)
Other	5(3)	6(4)
Admission Hunt-Hess Grade			0.014
I	86(48)	58(35)
II	33(18)	27(16)
III	37(20)	51(31)
IV	17(9)	23(14)
V	7(4)	6(4)
Fisher Group			0.016
1	14(8)	2(1)
2	86(48)	70(42)
3	66(36)	82(49)
4	14(8)	12(7)

Genotype frequencies, listed in Table 3, were in Hardy-Weinberg equilibrium. For the eNOS promoter, -786T>C SNP, 14% of subjects with vasospasm had the CC genotype, compared to 8% without vasospasm (p=0.15). Unadjusted OR for the CC vs. TT genotype was 2.10 (95% CI 0.97-4.53, p=0.06), while the CT vs. TT genotype was 1.02 (95%CI 0.64-1.65, p=0.90). Therefore, we used CC vs. any T genotype (CT or TT) for our multivariable models. There were no differences in genotype frequencies for the 894G>T SNP (p=0.60) or VNTR (p=0.87). The unadjusted OR for 894G>T and VNTR were 0.91 (95%CI 0.43-1.95, p=0.81) and 0.82 (95%CI 0.19-3.49, p=0.78) respectively. Similarly, there were no significant differences in any of the eNOS genotype frequencies between subjects with and without infarct (Table 3). No significant interactions between eNOS polymorphisms and clinical characteristics were observed (P<0.05).

Table 3. Genotype frequency of eNOS polymorphisms by primary outcome vasospasm and secondary outcome cerebral infarct.

	All	No Vasospasm	Vasospasm	p-value	No Infarct	Infarct	p-value
	n(%)	n(%)	n(%)		n(%)	n(%)
eNOS -786T>C				0.15			0.74
TT	167(53)	91(54)	76(50)		77(53)	41(56)
CT	119(37)	64(38)	55(36)		53(37)	23(32)
CC	33(10)	12(8)	21(14)		15(10)	9(12)
TOTALS	319	167	152		145	73
eNOS 894G>T (E298D)				0.60			0.25
GG	141(60)	77(57)	64(63)		84(59)	44(60)
GT	63(27)	39(29)	24(24)		43(30)	16(22)
TT	32(13)	19(14)	13(13)		16(11)	13(18)
TOTAL	236	135	101		143	73
eNOS VNTR*				0.87			0.65
bb	184(75)	107(76)	77(74)		114(77)	55(71)
ab	53(22)	29(20)	24(23)		29(20)	19(25)
aa	8(3)	5(4)	3(3)		5(3)	3(4)
TOTAL	245	141	104		148	77

^*a=4 allele, b=5 or 6 allele, as 6 is very rare occurring in only 6 cases

For -786T>C, the CC genotype compared to any T genotype was associated with a 2.97 increased odds of vasospasm (95%CI 1.32-6.67, p=0.008) after adjustment for age, gender, race/ethnicity, Hunt-Hess grade and Fisher group. We performed a subset analysis of Whites only, our largest ethnic subgroup, to minimize the confounding effect of population stratification. The association of CC genotype with vasospasm in 206 White subjects had a similar effect size (OR 2.09, 95%CI 0.88-4.98, p=0.09; Table 4).

Table 4. Association of eNOS -786T>C polymorphism with angiographic vasospasm.

	Unadjusted			Adjusted			Whites Only
	OR	95% CI	p-value	OR	95% CI	p-value	OR	95% CI	p-value
eNOS -786T>C(CC vs. any T)	2.09	0.99-4.40	0.05	2.97	1.32-6.67	0.008	2.09	0.88-4.98	0.09
Fisher Group 3	--	--	--	2.06	1.25-3.42	0.005	1.86	1.00-3.46	0.05
Female gender	--	--	--	1.30	0.77-2.21	0.32	1.36	0.70-2.64	0.36
Age, year	--	--	--	0.96	0.94-0.98	<0.001	0.95	0.93-0.98	<0.001
White vs. non-White	--	--	--	0.71	0.43-1.17	0.18	--	--	--
Hunt-Hess Grade <3	--	--	--	0.49	0.30-0.81	0.006	0.56	0.30-1.05	0.07

Discussion

We report the first quantitative estimation of the association of eNOS polymorphisms with risk of angiographic vasospasm after aneurysmal SAH. The CC genotype of the promoter -786T>C SNP was associated with nearly three-fold increased odds of angiographic vasospasm. The association of vasospasm with the CC genotype was consistent in all our logistic regression models and remained significant even after Bonferroni correction for multiple comparisons. This association was similar in our analysis of Whites only.

To date, the -786T>C SNP is the only eNOS polymorphism associated with cerebral vasospasm. Our results replicate the findings of the first study to report an association with this promoter SNP on 28 cases selected by CT criteria for the highest risk of vasospasm.¹⁶ The strengths of our study are a large prospective cohort and detailed radiographic outcome assessment.

We were able to show an association for the promoter polymorphism (-786T>C) with a relatively small sample size. For the -786T>C and 894G>T polymorphisms, our sample size could detect an effect size greater than 1.5-2.0. However, for the VNTR, we could only detect an effect size greater than 2-3.⁴⁶ We defined our outcome using the gold standard cerebral angiography for vasospasm, and included only cases that were eligible for at least one cerebral angiogram during the vasospasm risk period. Using this outcome measure would likely miss milder cases of vasospasm that responded to medical therapy alone and did not require angiography or interventional treatments. Excluding milder cases as having no vasospasm likely biases our results toward the null hypothesis. We did not find any association of eNOS genotype with cerebral infarct. Although this may also be a result of too small a sample size to detect an effect, progression to ischemia and infarct likely involves mechanisms distinct from NO pathways that may be involved in angiographic vasospasm.

The small sample size of ethnic minorities limited our ability to evaluate associations with eNOS polymorphisms within ethnic subgroups. There are clear differences in genotype frequency among ethnic groups.^47-49 Our minor allele frequencies were consistent with the NCBI dbSNP database (http://www.ncbi.nlm.nih.gov/SNP/) and previously published studies. ^{16, 47-49} Our strongest association was with the CC genotype (-786C>T) while previous studies reported the highest risk for CT heterozygotes.^{16, 22} In these previous studies among Japanese and Korean populations, CC genotypes occurred in less than 1% of subjects.^{47, 49} We need to confirm our findings in larger cohorts of non-White ethnic subgroups.

We hypothesize that eNOS polymorphisms contribute to individual variability in angiographic vasospasm, and our study supports the hypothesis that NO is important in the pathogenesis of angiographic vasospasm. Biological evidence from animals and humans suggests that differential expression of eNOS leads to decreased NO levels after SAH.⁵⁰ Specifically, the promoter -786T>C polymorphism has been associated with decreased eNOS promoter activity and lower nitrite/nitrate levels in subjects with coronary spasm.^{23, 25, 51}

Identifying patients at high risk for angiographic vasospasm and improving our ability to select patients for potentially harmful treatment has important clinical relevance. Vasospasm remains one of the most challenging management issues for clinicians since clinical and radiographic factors have been unreliable in predicting which patients will have significant clinical consequences from cerebral ischemia and infarct.⁴¹ Individual genetic variability in the response to SAH likely contributes to the clinical variability seen among subjects with similar presenting symptoms. Early initiation of treatment for vasospasm may prevent morbidity and mortality after aneurysm rupture and improve outcomes. Conversely, better selection of patients may prevent complications of aggressive therapy such as induced hypertension, hypervolemia, and interventional therapies. Genetic screening for high risk individuals may become clinically available in the near future. Future studies of the specific role of the promoter polymorphism (−786T>C) may also determine the effect of NO regulated by this pathway, and identify potential targets for treatment.