The negative effect of aging on cerebral venous outflow in acute ischemic stroke

Abstract

Cortical venous outflow (VO) represents an imaging biomarker of increasing interest in patients with acute ischemic stroke due to large vessel occlusion (AIS-LVO). We conducted a retrospective multicenter cohort study to investigate the effect of aging on VO. A total of 784 patients met the inclusion criteria. Cortical Vein Opacification Score (COVES) was used to assess VO profiles on admission CT angiography. Cerebral microperfusion was determined using the hypoperfusion intensity ratio (HIR) derived from perfusion imaging. Arterial collaterals were assessed using the Tan scale. Multivariable regression analysis was performed to identify independent determinants of VO, HIR and arterial collaterals. In multivariable regression, higher age correlated with worse VO (adjusted odds ratio [95% CI]; 0.83 [0.73–0.95]; P = 0.006) and poorer HIR (β coefficient [95% CI], 0.014 [0.005–0.024]; P = 0.002). The negative effect of higher age on VO was mediated by the extent of HIR (17.3%). We conclude that higher age was associated with worse VO in AIS-LVO, partially explained by the extent of HIR reflecting cerebral microperfusion. Our study underlines the need to assess collateral blood flow beyond the arterial system and provides valuable insights into deteriorated cerebral blood supply in elderly AIS-LVO patients.

Introduction

In times of demographic change and aging populations, it is of particular interest to gain a deeper pathophysiological understanding of how higher age influences ischemic stroke and its clinical course. There is much evidence that aging reduces the likelihood of favorable functional outcomes in patients with acute ischemic stroke due to large vessel occlusion (AIS-LVO), presumably due to higher comorbidity and lower capacity of rehabilitation in the older population.^{1 –4}

Another explanation for a larger stroke burden in the elderly might be a deterioration of collateral blood flow at a higher age. Arterial collateralization has been studied extensively in stroke research and was found to be associated with smaller infarct core volumes, higher rates of successful recanalization after mechanical thrombectomy and improved functional outcomes.^{5 –8} However, the effect of aging on arterial collateralization remains controversial. While several previous studies support a negative association between age and arterial collateralization,^{9 –13} others could not validate this finding.^5,8,14,15

Importantly, cerebral blood supply is not solely determined by arterial inflow, but also by microvascular circulation and venous drainage. The quality of cerebral microperfusion can be assessed by the hypoperfusion intensity ratio (HIR) derived from perfusion imaging. ¹⁶ Further downstream, cerebral venous outflow (VO) can serve as a surrogate for sustained blood transit through ischemic brain tissue. ¹⁷ A prior study found a strong correlation between HIR and VO profiles in AIS-LVO. ¹⁷ While age-related alterations in cerebral perfusion, for instance due to arteriolosclerosis, have been described extensively for the arterial component of the vascular circuit,^{18 –20} little is known about the age-related alterations of HIR or venous drainage in the setting of AIS-LVO. This study aims to investigate the effect of aging on cerebral collateralization at various vascular levels within the framework of AIS-LVO. We hypothesized that aging adversely affects the extent of cerebral VO due to a deterioration of HIR in elderly patients.

Materials and methods

Study design

This study reports against the STROBE guideline for observational studies. ²¹ In this retrospective multicenter cohort study, we included patients who were treated for AIS-LVO between October 1, 2013 and January 31, 2021 in two comprehensive stroke centers in Europe and the United States (University Medical Center Hamburg-Eppendorf, Germany and Stanford University School of Medicine, CA). Data on AIS-LVO patients were collected from the continuously maintained stroke database at each center. Baseline patient, imaging and treatment characteristics were obtained from medical records and diagnostic imaging.

The patient inclusion criteria were as follows: [1] Triage for mechanical thrombectomy within 16 hours after symptom onset; [2] multimodal computed tomography with pretreatment non-contrast head computed tomography, single-phase computed tomography angiography and computed tomography perfusion and [3] acute ischemic stroke due to anterior circulation large vessel occlusion of the internal carotid artery or M1 or M2 segment of the middle cerebral artery. Supplementary Figure S1 details a flowchart of patient inclusion and exclusion criteria. Of the 813 patients who underwent triage for thrombectomy, 784 patients were included.

This study was approved by the institutional review boards at each of the included centers (University Medical Center Hamburg-Eppendorf, Germany and Stanford University School of Medicine, CA). The study was conducted in accordance with the ethical guidelines of the local ethics committees and in accordance with the Declaration of Helsinki and the Health Insurance Portability and Accountability Act (HIPAA). Informed consent was waived by the institutional review boards due to the retrospective design of the study.

Assessment of arterial collateralization

Arterial collateralization was assessed on admission computed tomography angiography using the modified Tan scale. ²² Favorable arterial collaterals were defined as filling of ≥50% of the middle cerebral artery territory on the affected hemisphere. Ratings were performed by two experienced neuroradiologists (T.D.F. and J.J.H., with 11 and 16 years of working experience).

Assessment of perfusion parameters

Computed tomography perfusion imaging was analyzed using the software platform RAPID (iSchemaView, Menlo Park, California, USA). Baseline ischemic core volume was identified by RAPID as tissue with a relative cerebral blood flow of <30% compared to the mean cerebral blood flow measured in regions of both hemispheres that do not have Tmax delays as the control. Penumbra volume was defined as the difference between the brain volume with Tmax >6 s minus the baseline ischemic core volume. According to previous studies, HIR was defined as Tmax >10 s volume (severe hypoperfusion) over Tmax >6 s volume (critical hypoperfusion).^{23 –25} Lower HIR values indicate better cerebral microperfusion within the ischemic brain volume.

Assessment of cortical venous outflow

Cortical VO profiles were assessed on admission computed tomography angiography using the Cortical Vein Opacification Score (COVES). ²⁶ COVES reflects the main venous drainage pathways of the middle cerebral artery territory by grading the opacification of three cortical veins on a scale from 0 to 2 points (0: not visible, 1: moderate opacification, 2: full opacification). The cumulative COVES ranges from 0 to 6 points. Favorable VO was defined as COVES ≥ 3. ²⁷ Ratings were performed by two neuroradiologists with substantial inter-reader agreement (T.D.F. and J.J.H., with 11 and 16 years of working experience). Discrepancies were settled by consensus readings.

Imaging analysis

The angiographic degree of recanalization was assessed by two experienced neuroradiologists (T.D.F. and J.J.H., with 11 and 16 years of working experience), using the Thrombolysis In Cerebral Infarction (TICI) scale on final digital subtraction angiography images. Discrepancies were settled by consensus readings.

Assessment of functional outcome

The modified Rankin Scale scores at 90 days were determined by a stroke neurologist or registered study nurse in a telephone or face-to-face assessment.

Outcome measures

Favorable VO was defined as the primary outcome measure. Favorable arterial collaterals and HIR were defined as secondary outcome measures.

Statistical analysis

All analyses were performed using R statistical software (version 4.1.2, R Project for Statistical Computing), RStudio statistical software (version 2021.09.1 + 372, Rstudio) and Stata/MP statistical software (version 17.0, StataCorp). A two-tailed p-value of <0.05 was considered significant for all statistical tests.

We performed univariable comparisons between VO+ and VO− patients (Table 1). Kolmogorov-Smirnov tests were used to test data distributions for normality. Continuous variables were reported as median and interquartile range (IQR). Categorical variables are described as counts and percentages. Continuous and categorical variables were compared between groups using Mann-Whitney U test and chi-squared test, respectively.

Table 1.

Patient characteristics, imaging findings, treatment characteristics and clinical outcomes dichotomized by venous outflow (VO).

	All patients(n = 784)	VO−(n = 475)	VO+(n = 309)	P value a
Patient characteristics
Age (years), median (IQR)	76 (64–83)	77 (66–84)	72 (62–81)	<0.001 (1)
Male sex, n (%)	383 (48.9%)	212 (44.6%)	171 (55.3%)	0.003 (2)
Atrial fibrillation, n (%)	310 (39.8%)	195 (41.2%)	115 (37.7%)	0.327 (2)
Hypertension, n (%)	540 (69.2%)	329 (69.6%)	211 (68.7%)	0.807 (2)
Diabetes, n (%)	172 (22.1%)	105 (22.2%)	67 (21.8%)	0.890 (2)
Blood glucose on admission (mg/dL), median (IQR)	122 (105–148)	123 (106–152)	119 (103–144)	0.035 (1)
Lipid disorder, n (%)	223 (31.9%)	129 (30.5%)	94 (34.1%)	0.323 (2)
Admission NIHSS, median (IQR)	15 (9–19)	17 (12–20)	10 (7–16)	<0.001 (1)
Imaging characteristics
Time from symptom onset to imaging (min), median (IQR)	180 (92–350)	186 (99–410)	173 (86–320)	0.087 (1)
Location of arterial occlusion				<0.001 (2)
ICA, n (%)	155 (19.8%)	123 (25.9%)	32 (10.4%)
MCA – M1, n (%)	471 (60.1%)	297 (62.5%)	174 (56.3%)
MCA – M2, n (%)	158 (20.2%)	55 (11.6%)	103 (33.3%)
ASPECTS, median (IQR)	8 (6–9)	7 (6–9)	9 (7–10)	<0.001 (1)
Penumbra volume (ml), median (IQR)	90 (49–138)	98 (56–145)	77 (45–124)	<0.001 (1)
Baseline ischemic core volume (ml), Median (IQR)	9 (0–31)	17 (3–45)	4 (0–13)	<0.001 (1)
Favorable arterial collaterals (Tan scale), n (%)	531 (67.7%)	257 (54.1%)	274 (88.7%)	<0.001 (2)
HIR, median (IQR)	0.5 (0.3–0.6)	0.5 (0.4–0.6)	0.3 (0.2–0.5)	<0.001 (1)
Treatment characteristics
Administration of tPA, n (%)	387 (50.1%)	195 (41.8%)	192 (62.7%)	<0.001 (2)
Mechanical thrombectomy, n (%)	695 (88.6%)	422 (88.8%)	273 (88.3%)	0.832 (2)
Successful recanalization [TICI 2 b/2c/3], n (%)	559 (80.4%)	319 (75.6%)	240 (87.9%)	<0.001 (2)
Clinical outcomes
24-hour NIHSS, median (IQR)	12 (5–19)	16 (9–21)	6 (2–11)	<0.001 (1)
mRS score at 90-d follow-up, median (IQR)	4 (1–5)	5 (4–6)	2 (1–3)	<0.001 (1)

NIHSS: National Institutes of Health Stroke Scale; ICA: internal carotid artery; MCA: middle cerebral artery; ASPECTS: The Alberta Stroke Program Early CT score; HIR: hypoperfusion intensity ratio; TICI: thrombolysis in cerebral infarction.

^aCharacteristics were compared between 475 VO− and 309 VO+ patients with the use of either Mann-Whitney U test (1) for continuous variables or a chi-square test (2) for categorical variables. Statistical significance: p < 0.05.

Separate multivariable logistic regression models were designed to identify independent variables associated with [1] favorable arterial collaterals and [2] favorable VO. Multivariable linear regression was used to determine factors associated with HIR. Independent variables were included into the regression models after a priori selection based on prior studies. We conducted complete-case analysis for all regression models (for a detailed description of missing data points see Supplementary Figure 1). Adjusted odds ratios (aOR) with p-values and 95% confidence intervals (CI) were reported for each independent variable. We calculated the variance of inflation factor for each independent variable to exclude multicollinearity in the regression models. A mediation analysis was performed to investigate how HIR impacts the effect of aging on cortical VO (see Figure 1). We tested the significance of the indirect effect using bootstrapping.

Figure 1.

Effect of Aging on Venous Outflow is Partially Mediated by HIR. A mediation analysis was conducted to investigate the effect of aging on cortical venous outflow with mediation through the hypoperfusion intensity ratio (HIR), which approximates the extent of cerebral microperfusion. The relationship between aging and HIR is expressed by the regression coefficient (a), while (b) shows the relationship between HIR and venous outflow and (c′) the effect of aging on VO. The regression analyses of the total effect [c = (−0.04); P = 0.002], direct effect [c′ = −0.03; P = 0.008] and indirect effect [ab = (0.01)*(−0.69) = −0.0069; P < 0.001] were significant. Thus, HIR partially mediated the relationship between aging and venous outflow.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Results

Patient characteristics

A total of 784 AIS-LVO patients met the inclusion criteria (Supplemental Figure 1). The median age was 76 years and the sex ratio was balanced with 51.1% being women. Favorable arterial collaterals were found in 531 (67.7%) patients, while favorable VO was observed in 309 (39.4%) patients. The median HIR was 0.5 (IQR, 0.3–0.6). Successful mechanical recanalization (Thrombolysis in Cerebral Infarction 2 b, 2c or 3) was achieved in 559 of 695 patients (80.4%) and the median modified Rankin Scale Score at 90 days was 4 (IQR, 1–5 points). Please refer to Table 1 for more detailed information about baseline, imaging, and treatment characteristics.

Patient characteristics stratified by venous outflow

Patients with favorable venous outflow (VO+) were younger (median, 72 vs 77 years; P < 0.001) and more likely to be men (55.3% vs. 44.6%; P = 0.003). VO+ patients had higher admission ASPECTS (median, 9 vs 7 points; P < 0.001), higher rates of favorable arterial collaterals (88.7% vs 54.1%; P < 0.001) and lower HIR values compared with VO− patients (median, 0.3 vs 0.5; P < 0.001). In addition, modified Rankin Scale scores at 90 days were significantly lower in VO+ patients compared with VO− patients (median, 2 vs 5 points; P = 0.001).

Multivariable regression

We performed multivariable logistic regression analysis to identify independent determinants of favorable VO in AIS-LVO patients. Higher age was associated with lower odds of achieving favorable VO (aOR [95% CI], 0.83 [0.73–0.95]; P = 0.006). Moreover, favorable arterial collaterals (aOR [95% CI], 3.26 [2.07–5.23]; P < 0.001) increased, while higher HIR (aOR [95% CI], 0.88 [0.80–0.98]; P = 0.016) decreased the odds of favorable VO. Further significant independent variables were male sex, ASPECTS, baseline ischemic core volume and vessel occlusion site (see Table 2).

Table 2.

Multivariable logistic regression to predict VO on admission CTA in AIS-LVO patients.

	Dependent variable: Favorable VO
Independent variables	Adjusted odds ratio[95 % CI]	P value
Age (per 10 years)	0.83 [0.73–0.95]	0.006
Sex (male)	1.76 [1.19–2.61]	0.005
Blood glucose on admission (per 10 mg/dl)	0.98 [0.94–1.02]	0.347
Proximal vessel occlusion [ICA/M1] (yes)	0.15 [0.10–0.22]	<0.001
ASPECTS (per 1 point)	1.16 [1.04–1.29]	0.009
Favorable arterial collaterals (yes)	3.26 [2.07–5.23]	<0.001
Penumbra volume (per 10 ml)	0.99 [0.96–1.01]	0.220
Baseline ischemic core volume (per 10 ml)	0.87 [0.79–0.95]	0.004
HIR value (per 0.1 points)	0.88 [0.80–0.98]	0.016

n = 744 patients included. Statistical significance: p < 0.05.

ICA: internal carotid artery; MCA: middle cerebral artery; ASPECTS: The Alberta Stroke Program Early CT score; HIR: hypoperfusion intensity ratio.

Multivariable regression models to identify predictors of favorable arterial collaterals and HIR are displayed in Table 3. There was no significant association between age and the likelihood of favorable arterial collaterals after adjustment for covariates (aOR [95% CI], 1.04 [0.92–1.18]; P = 0.537). However, older age was linked to higher HIR values (β coefficient [95% CI], 0.014 [0.005–0.024]; P = 0.002). In addition, smaller baseline ischemic core volume (β coefficient [95% CI], 0.030 [0.026–0.035]; P < 0.001), and smaller penumbra volume (β coefficient [95% CI], 0.002 [0.000–0.004], P = 0.005) were associated with lower HIR.

Table 3.

Multivariable logistic regression model to predict favorable arterial collaterals and Multivariable linear regression to predict the extent of hypoperfusion intensity ratio (HIR).

	Dependent variable: Favorable arterial collaterals		Dependent variable: HIR (linear regression)
Independent variables	Adjusted Odds Ratio [95% CI]	P value	β	95% CI	P value
Age (per 10 years)	1.04 [0.92–1.18]	0.537	0.014	0.005 to 0.024	0.002
Sex (male)	1.27 [0.88–1.88]	0.200	0.047	0.020 to 0.074	<0.001
Blood Glucose on admission (per 10 mg/dl)	0.96 [0.93–1.00]	0.039	−0.003	−0.005 to 0.000	0.018
Proximal vessel occlusion [ICA/M1] (yes)	0.79 [0.53–1.19]	0.259	0.007	−0.022 to 0.036	0.647
ASPECTS (per 1 point)	1.10 [1.00–1.21]	0.048	−0.013	−0.020 to −0.005	<0.001
Favorable arterial collaterals (yes)	–	–	−0.045	−0.076 to −0.015	0.004
Penumbra volume (per ml)	0.99 [0.97–1.02]	0.511	0.002	0.000 to 0.004	0.005
Baseline ischemic core volume (per ml)	0.91 [0.85–0.97]	0.003	0.030	0.026 to 0.035	<0.001
HIR value (per 0.1 points)	0.85 [0.77–0.94]	0.002	–	–	–
Favorable VO (yes)	3.37 [2.14–5.41]	<0.001	−0.055	−0.087 to −0.023	<0.001

Favorable arterial collaterals: n = 744 patients included. HIR: n = 744 patients included.

Statistical significance: p < 0.05

ICA: internal carotid artery; MCA: middle cerebral artery; ASPECTS: The Alberta Stroke Program Early CT score; HIR: hypoperfusion intensity ratio.

Mediation analysis

We performed a mediation analysis to further investigate the influence of aging on cortical VO (Figure 1). The effect of aging on VO was partially mediated by the extent of HIR. The regression coefficients of the total effect [c = (−0.04); P = 0.002], direct effect [c′ = −0.03; P = 0.008] and indirect effect [ab = (0.01)*(−0.69) = −0.0069; P < 0.001] were significant. A total of 17.3% of the total effect of aging on VO was explained by HIR reflecting the extent of cerebral microperfusion.

Discussion

In this study, we analyzed the impact of aging on the cerebral collateral status of AIS-LVO patients on the arterial, microcirculatory, and venous level. We found that higher age was associated with lower odds of exhibiting favorable cerebral VO and HIR. The effect of aging on cortical VO was partially mediated by HIR. Notably, there was no significant association between higher age and arterial collateral status.

Prior studies suggest that higher age has a negative effect on arterial collateral status in acute ischemic stroke.^{9 –13} For instance, Wiegers et al. examined the MR CLEAN registry and found that poor arterial collateralization was more frequent at higher age. ¹² This finding was supported by research in rodents by providing evidence of a rarefication, an increase of tortuosity and higher vascular resistance of arterial collaterals in older subjects.^{28 –30} In contrast, other studies could not find such a detrimental effect of aging on arterial collaterals, which is in line with our data.^5,8,14,15 One possible explanation for these differences might be inherent to the design of well-established arterial collateral scores. The mere opacification of macrovascular collateral arteries might be insensitive to age-related changes in collateral flow, but represents the main feature captured by CTA-based arterial collateral scores. However, the negative effect of aging on the microvascular circulation is not likely to be reflected by arterial collateral scores. ²⁰

Cortical VO might better account for age-related changes of the microvascular circulation by serving as an indirect surrogate for sustained blood flow through ischemic brain tissue. Accordingly, we found that higher age was associated with a lower likelihood of favorable VO in acute ischemic stroke and that this effect was partially mediated by HIR reflecting cerebral microperfusion. Previous studies have shown that the physiological process of aging intrinsically results in a decreased cerebral tissue perfusion (e.g. through lower cardiac output, cerebral atrophy and altered cerebral metabolism),^{31 –33,39} which may have an impact on cortical VO in elderly patients. However, our study does not include data from healthy probands, but from stroke patients in the acute stroke setting, who can be assumed to have impaired cerebral tissue perfusion beyond the level of physiological aging. Further research is needed to assess the effect of aging on VO in a sample of healthy participants.

An additional factor that may lead to a deterioration of microcirculation in elderly patients may be explained by the presence of cerebral small vessel disease. Cerebral small vessel disease represents the most common vascular disease in the elderly which affects arterioles, capillaries and venules alike. ³⁴ Interestingly, a retrospective study observed an association between chronic cerebral small vessel disease and poor recruitment of arterial collaterals in AIS-LVO patients. ³⁵ Cerebral small vessel disease and its associated age-related risk factors including arteriolosclerosis and chronic vessel wall inflammation may further reduce the microvascular blood transit through brain tissue and its downstream cortical VO.^18,20,36,37

The indirect deterioration of cortical VO via higher (poorer) HIR might be reinforced by direct changes to cerebral veins related to aging. Recent studies highlighted the role of age-related alterations of the cerebral venous system in the pathogenesis of diseases of the central nervous system such as leukoaraiosis and vascular cognitive impairment.^34,38 Pathophysiological findings, for example venous collagenosis, venular tortuosity and increased stiffness of the venous wall, might diminish cortical VO at higher age. However, our study does not provide any evidence supporting the association between aging, structural changes in cerebral veins and consecutively impaired VO in AIS-LVO patients. Further studies are required to identify potential pathological correlates of altered VO in acute ischemic stroke.

Limitations

Our study has several limitations. First, the retrospective study design may lead to selection bias and reduce the generalizability of our findings. Second, complete-case analysis with exclusion of patients with missing data might introduce further bias. Third, VO assessment on single-phase computed tomography angiography might be affected by the selected imaging protocol, including acquisition timing and rate of contrast bolus injection. ²⁶ Fourth, our data did not include well established measures for chronic small vessel ischemia, e.g. the Fazekas scale, which may influence the extent of HIR and VO profiles.

Conclusion

This is the first study which investigates the relationship between aging and collateral status in AIS-LVO on the arterial, microvascular, and venous level. Higher age was associated with poorer cortical VO and worse HIR in acute ischemic stroke. The effect of aging on cortical VO was partially mediated by the extent of HIR, which reflects microvascular perfusion in the ischemic region. Our study underlines the need to assess collateral blood flow beyond the arterial system and provides valuable insights into deteriorated cerebral blood supply in elderly AIS-LVO patients.