Determinants of Leptomeningeal Collateral Status in Acute Ischemic Stroke: A Systematic Review and Meta‐Analysis of Observational Studies Authors: Kunyi Li, Hua Jiang, Jianping Yu, Yong Liu, Lili Zhang, Bi Ma, Shu Zhu, Yinkuang Qi, Shuang Li, Yan Huang, Yuhan Yang, Xun Xia, Lan Wen Published: 2024 ABSTRACT AbstractBackgroundLeptomeningeal collateral status is a major determinant of outcomes in patients with acute ischemic stroke; however, the factors that determine collateral status are not well understood. We conducted a comprehensive systematic review and meta‐analysis to identify determinants associated with collateral status in patients with anterior circulation infarction.Methods and ResultsThe PubMed, EMBASE, Web of Science, and Cochrane Central Register of Controlled Trials databases were searched for studies that reported the determinants of leptomeningeal collateral status in acute ischemic stroke between January 2000 and June 2023. A random‐effects meta‐analysis model was used to pool the determinants of leptomeningeal collateral status. Eighty‐one studies with 17 366 patients met the inclusion criteria. We analyzed 31 potential risk factors, and the results indicated that worse leptomeningeal collateral status was significantly associated with older age (weighted mean difference, 1.22 [95% CI, 0.69 to 1.76]), male sex (odds ratio [OR], 1.12 [95% CI, 1.02 to 1.23]), hypertension (OR, 1.27 [95% CI, 1.15 to 1.40]), diabetes (OR, 1.21 [95% CI, 1.10 to 1.33]), atrial fibrillation (OR, 1.26 [95% CI, 1.09 to 1.46]), cardioembolic stroke (OR, 1.27 [95% CI, 1.04 to 1.55]), internal carotid artery occlusion (OR, 1.84 [95% CI, 1.50 to 2.25]), and higher admission blood glucose (weighted mean difference, 8.74 [95% CI, 2.52 to 18.51]).ConclusionsHypertension and diabetes could be modifiable risk factors associated with leptomeningeal collateral status. Older age and male sex could be nonmodified risk factors. Further high‐quality therapeutic studies focusing on controlling risk factors are needed to support our findings.Keywords: collateral status, meta, risk factor, strokeSubject Categories: Ischemic Stroke Clinical Perspective What Is New? This systematic review and meta‐analysis found that hypertension and diabetes could be modifiable risk factors associated with leptomeningeal collateral status. This systematic review and meta‐analysis found that hypertension and diabetes could be modifiable risk factors associated with leptomeningeal collateral status. Older age and male sex could be nonmodified risk factors. Older age and male sex could be nonmodified risk factors. What Are the Clinical Implications? As several risk factors of leptomeningeal collateral status reported in this study could be partly explained by endothelial dysfunction, ameliorating endothelial dysfunction might be a promising therapeutic target. As several risk factors of leptomeningeal collateral status reported in this study could be partly explained by endothelial dysfunction, ameliorating endothelial dysfunction might be a promising therapeutic target. Leptomeningeal arterial collaterals are preexisting anastomoses that cross‐connect a small number of distal arterioles of the cerebral arteries. When the proximal artery is occluded in acute ischemic stroke, leptomeningeal collaterals are compensatory pathways that redistribute flow and protect tissue at risk from ischemia. 1 Several studies indicate that collateral status is a major determinant of outcomes in patients with ischemic stroke. 2 , 3 Patients with poor collateral status have larger infarctions and increased rates of hemorrhagic transformation. 4 , 5 Patients with favorable collateral status have more salvageable brain tissue. A favorable collateral status has also been associated with greater benefit of intravenous thrombolysis and endovascular therapy. 6 Moreover, treatment decisions for the reperfusion therapies are increasingly guided by the evaluation of collateral status. 7 Therefore, identifying the potential risk factors of collateral status could help to find ways to improve collateral status and thus increase chances of better clinical outcomes. Recently, numerous studies have shown that many factors, including advanced age, hypertension, diabetes, smoking, and hyperlipidemia are associated with poor collateral status. 8 However, most of these studies only evaluate some potential risk factors with small sample sizes. The findings of these studies were often inconsistent. Thus, we performed this comprehensive systematic review and meta‐analysis to identify the risk factors associated with collateral status in anterior circulation stroke. METHODS This systematic review and meta‐analysis was reported based on the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guidelines 9 (Table S1). The review protocol has been registered at the International Prospective Register of Systematic Reviews (CRD 42022359378). No institutional review board approval or informed consent was required for conducting this systematic review and meta‐analysis. The data that support the findings of this study are available from the corresponding author on reasonable request. Literature Search Strategy A systematic literature search was conducted in PubMed, EMBASE, Web of Science databases, and Cochrane Central Register of Controlled Trials databases for articles published between January 1, 2000, and June 1, 2023. The keywords searched were stroke, acute cerebrovascular accident, and collateral circulation. A detailed search strategy is described in Data S1. In addition, the reference lists of eligible articles were manually screened for other eligible articles to avoid missing any relevant studies. The authors of the articles were contacted by email for detailed information if necessary. Two independent investigators (S.L. and H.J.) evaluated the results and resolved any disagreement through discussions with a third investigator (Y.H.). Eligibility Criteria and Data Extraction We used a modification of the Population, Intervention, Comparator, Outcome scheme for eligibility criteria. Studies were eligible if they met all of the following inclusion criteria: (1) Population: patients with acute ischemic stroke in anterior circulation; (2) Intervention: imaging methods including digital subtraction angiography and computer tomography angiography; (3) Comparator: collateral circulation status; and (4) Outcome: factors associated with collateral circulation status, evaluated by digital subtraction angiography or computer tomography angiography before the initiation of intravenous alteplase or endovascular therapy. Either the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology, Tan, or comparable collateral grading method was used to define “good” and “poor” status according to the included studies. We excluded case reports, case series, letters to the editor, and conference abstracts. Two reviewers (Y.L. and J.P.Y.) independently extracted the data from each article. The results of the extracted data were cross‐checked to confirm the accuracy of the data. Disagreements were resolved by discussion with a third reviewer (L.L.Z.). The following information was recorded: author name, publication year, study location, sample size, and factors associated with collateral circulation status. If relevant information was missing, the corresponding author was contacted. Quality Assessment The Newcastle‐Ottawa Scale tool was used to evaluate the risk bias of observational studies. The Newcastle‐Ottawa Scale tool evaluates the methodological quality of cohort studies through 3 dimensions: selection criteria, comparability, and outcome. 10 Newcastle‐Ottawa Scale scores are summed with a maximum of 9 points, as recommended by the Cochrane Non‐Randomized Studies Methods Working Group. A score of ≥7 suggested high quality, 4 to 6 indicated moderate quality, and a score of ≤3 indicated low quality. Two study reviewers (B.M. and S.Z.) independently assessed the methodological quality of the included studies. Any disagreements were resolved through discussion with a third study reviewer (Y.K.Q.). Certainty of Evidence The risk factors in each relevant article were assessed, and the strength of evidence supporting these variables was determined using the Grading of Recommendations Assessment, Development, and Evaluation system. 11 According to this system, the quality of evidence is assessed based on factors including study design, risk of bias, imprecision, inconsistency, indirectness, publication bias, and effect size. There are 4 levels of evidence: high, moderate, low, and very low. Statistical Analysis All statistical analyses were performed using STATA version 14.0 (STATA, College Station, TX). Odds ratio (OR) was used to compare the categorical factor data, and weighted mean difference was used to compare the continuous factor data. Both fixed‐effect and random‐effect models were used in analyses. Heterogeneity was tested using Cochran's Q statistic, and a P value <0.1 indicated significant heterogeneity. The extent of heterogeneity was quantified using the I 2 statistic. 12 Weighted SD was also measured to investigate the heterogeneity more sufficiently. For the meta‐analysis of the factors such as male sex and age, a “leave‐one‐out” sensitivity analysis was used to assess the robustness of the pooled results. 13 Publication bias was checked by visual inspection of funnel plots and tested for significance using Egger's regression test and Begg's rank correlation test for funnel plot asymmetry. 14 , 15 A subgroup analysis was performed according to number of patients, location and imaging method. Generally, only factors that were investigated in at least 3 studies were used. All meta‐analyses were conducted on univariate data and should be interpreted as unadjusted pooled estimates. All P values were 2‐sided, and a P value <0.05 was considered to indicate statistical significance. RESULTS Study Selection and Characteristics The flow chart of study selection is shown in the Figure. For potentially eligible studies published between January 1, 2000, and June 1, 2023, 5302 records were identified after duplicates were removed. Then, 5139 records were excluded by reviewing the titles and abstracts, and another 82 records were excluded by reviewing the full text. Flow chart of the included studies. Overall, 81 2 , 4 , 7 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 , 68 , 69 , 70 , 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 , 84 , 85 , 86 , 87 , 88 , 89 , 90 , 91 , 92 , 93 primary studies comprising 17 366 patients were included in the systematic review, of which 29 were prospective studies, 52 were retrospective studies, 18 were multicenter studies, and 63 were single‐center studies. Of these studies, 18 were in North America (22.2%), 25 in Europe (30.8%), 33 in Asia (40.7%), 4 in Oceania (4.9%), and 1 in mixed regions. A total of 48.6% (7989/16453) of patients were men, and the mean age was 68.9 years. Pretreatment collateral status was assessed by digital subtraction angiography and computer tomography angiography in 34 and 47 studies, respectively. The detailed characteristics of the included studies are shown in Table S2. Quality Assessment We clarify the study types by using the Population, Intervention, Comparator, Outcome scheme. All the included studies were observational studies; therefore, the Newcastle‐Ottawa Scale tool was used to evaluate the quality of the included studies and is presented in Table S2. Certainty of Evidence The quality of the evidence was assessed by Grading of Recommendations Assessment, Development, and Evaluation guidelines. Observational studies started with a “low quality” rating, and the quality of evidence was downgraded for “imprecision.” Thus, the overall quality of the evidence was rated as “very low.” Predictors of Collateral Status We performed a meta‐analysis to investigate the factors associated with collateral status in anterior circulation infarction (Table, Figure S1). The results showed that older age, male sex, high National Institutes of Health Stroke Scale score, hypertension, diabetes, atrial fibrillation, high admission blood glucose, cardioembolic stroke, and internal carotid artery occlusion were significantly associated with worse collateral status (Table, Figure S1). The “leave‐one‐out” sensitivity analysis and the test for publication bias are shown in Figures S2 and S3. Table S3 summarizes the subgroup analyses of relevant factors associated with collateral status. Table 1Relevant Factors Associated With Pretreatment Collateral StatusRelevant factorsSystematic review or meta‐analysisHeterogeneityWeighted SDEgger's testBegg's testNumberOdds ratio/Weighted mean differenceStudiesPatients(95% CI) fixed model P value(95% CI) random mModel P value I 2, % P value P value P valueGeneral characteristicsAge*48104751.22 (0.69 to 1.76)<0.0011.42 (0.75 to 2.08)<0.001200.1172.070.1660.783Male sex† 74163081.16 (1.09 to 1.24)<0.0011.12 (1.02 to 1.23)0.01430.90.0070.450.1710.508Body mass index, kg/m2† 56210.47 (−0.22 to 1.16)0.1800.47 (−0.22 to 1.16)0.1800.5930.640.3540.462Baseline National Institutes of Health Stroke Scale score† 1212502.44 (1.88 to 3.00)<0.0012.62 (1.46 to 3.77)<0.00173.6<0.0011.960.6040.15Systolic BP at admission, mm Hg† 1864301.53 (0.24 to 2.81)0.0201.61 (−0.02 to 3.25)0.05215.50.2693.020.4860.495Diastolic BP at admission, mm Hg† 1655800.52 (−0.31 to 1.34)0.218−0.13 (−1.59 to 1.32)0.85951.50.0092.020.218>0.99Clinical historyHypertension† 61147451.23 (1.15 to 1.32)<0.0011.27 (1.15 to 1.40)<0.00128.60.0220.550.1030.271Diabetes† 60142771.19 (1.09 to 1.30)<0.0011.21 (1.10 to 1.33)<0.0017.40.3140.490.4180.425Hyperlipidemia† 50125531.10 (1.01 to 1.19)0.0271.1 (0.99 to 1.21)0.07815.20.1830.510.8140.776Atrial fibrillation† 4199381.24 (1.14 to 1.36)<0.0011.26 (1.09 to 1.46)0.00251.6<0.0010.650.9360.814Previous stroke† 2283601.15 (1.02 to 1.30)0.0271.19 (0.997 to 1.42)0.053250.140.480.1820.338Previous transient ischemic attack† 35330.84 (0.40 to 1.77)0.6520.89 (0.41 to 1.94)0.7635.20.3480.550.782>0.99Heart failure† 710940.85 (0.61 to 1.19)0.3470.86 (0.62 to 1.21)0.38900.5360.370.6950.368Peripheral artery disease† 321371.47 (1.07 to 2.01)0.0171.42 (0.88 to 2.3)0.15512.50.3190.280.4340.296Current smoking† 44107361.07 (0.97 to 1.17)0.1761.05 (0.94 to 1.18)0.412160.1830.510.6270.413Alcohol consumption† 55251.26 (0.85 to 1.85)0.2451.26 (0.86 to 1.86)0.24300.8840.240.7690.806Laboratory data at admissionLeukocytes, per μL*32610.5 (−0.48 to 1.48)0.3160.5 (−0.48 to 1.48)0.31600.7980.530.836>0.99Platelets, ×10 000/μL*3454−3.06 (−16.26 to 10.15)0.650−3.06 (−16.26 to 10.15)0.6500.9811.330.503>0.99Hematocrit, %*33610.49 (−0.83 to 1.81)0.4650.23 (−1.77 to 2.23)0.82540.1141.050.0810.296Blood glucose, mg/dL*1116698.74 (4.15 to 13.33)<0.0019.10 (1.60 to 16.54)0.01655.40.01311.360.7410.533Glycosylated hemoglobin, %*33580.15 (−0.07 to 0.36)0.1710.12 (−0.32 to 0.56)0.58858.30.0910.370.906>0.99Total cholesterol, mmol/L*58140.01 (−0.13 to 0.16)0.8700.1 (−0.17 to 0.36)0.46756.20.0580.230.2410.806Triglyceride, mmol/L*4745−0.07 (−0.2 to 0.06)0.275−0.04 (−0.23 to 0.14)0.64341.70.1610.170.3750.734Low‐density lipoprotein, mmol/L*61010−0.02 (−0.11 to 0.08)0.724−0.005 (−0.12 to 0.11)0.93521.30.2730.120.240.26High‐density lipoprotein, mmol/L*47450.03 (−0.02 to 0.07)0.2840.03 (−0.01 to 0.15)0.66884.4<0.0010.100.9860.734Stroke typeCardioembolic stroke† 2151971.12 (1.00 to 1.26)0.0421.27 (1.04 to 1.55)0.02148.20.0070.650.0930.651Large‐artery atherosclerosis† 1624030.79 (0.66 to 0.96)0.0150.76 (0.59 to 0.99)0.03929.70.1260.410.5070.62Complete circle of Willis† 49940.73 (0.53 to 1.00)0.0490.74 (0.52 to 1.05)0.09511.70.3340.210.8290.734Occlusion locationInternal carotid artery† 38101081.77 (1.60 to 1.95)<0.0011.84 (1.50 to 2.25)<0.00167.8<0.0011.560.4470.365MCA‐M1† 3199110.82 (0.75 to 0.89)<0.0010.82 (0.66 to 1.03)0.08280.5<0.0010.480.7910.377MCA‐M2† 2378260.77 (0.68 to 0.87)<0.0010.65 (0.47 to 0.9)0.0178.1<0.0010.540.1220.916The data effect size of the risk factors marked with “*” is weighted mean difference (95% Cl). The data effect size of the risk factors marked with “†” is odds ratio (95% Cl). BP indicates blood pressure; and MCA, middle cerebral artery. Relevant Factors Associated With Pretreatment Collateral Status The data effect size of the risk factors marked with “*” is weighted mean difference (95% Cl). The data effect size of the risk factors marked with “†” is odds ratio (95% Cl). BP indicates blood pressure; and MCA, middle cerebral artery. DISCUSSION In this meta‐analysis with 17 366 patients, we investigated 31 potential risk factors that are associated with poor collateral status in anterior circulation stroke. We found that worse collateral status was associated with the following clinical and imaging determinants: older age, male, hypertension, diabetes, cardioembolic stroke, internal carotid artery occlusion, and higher admission glucose level. It is not surprising that older age could be associated with poor collateral status. A possible explanation has been proposed that aging could cause dysfunction of the endothelium‐dependent vasodilatory responses, leading to leptomeningeal collateral rarefaction with a decrease in collateral diameter and density. 94 , 95 In addition, several studies have suggested that the association between age and collateral status could be mediated by many other age‐related factors, such as hypertension and diabetes. 20 , 81 We also found worse collateral status in men, which may partly be explained by different sex hormones. Several experimental and clinical studies have suggested that ovarian steroids can reverse endothelial dysfunction and stimulate collateral vessel formation in ischemic tissues. 96 , 97 Although demographic characteristics are nonmodifiable risk factors, these data could still be useful in the selection of ideal patients for collateral augmentation treatment. Our study suggests that hypertension is significantly associated with poor collateral status. In animal experiments, spontaneously hypertensive rats had poor cerebral collaterals due to impaired leptomeningeal collateral growth, suggesting a major role of hypertension in reducing leptomeningeal collateral flow. 98 Notably, the pharmacological actions of antihypertensive drugs may improve collateral status. An animal experiment indicated that an angiotensin‐converting enzyme inhibitor could reverse vasoconstriction and impaired dilation of pial collaterals in a chronic hypertension rat model, which could rescue unfavorable collateral status. 99 Moreover, previous studies have shown that other antihypertensive drugs such as calcium channel antagonists and β blockers could alleviate endothelial dysfunction in patients with chronic hypertension. 100 , 101 , 102 These findings suggest that chronic hypertension is associated with unfavorable collateral status. Furthermore, some antihypertensive drugs may contribute to better collateral status thus improving outcomes in patients with acute ischemic stroke, which should be evaluated in further studies. Our study suggested that the findings of the association between admission blood pressure (BP) and collateral status were inconsistent. A previous retrospective study found that admission BP may be associated with decreased collateral status, as assessed by computed tomography perfusion. 103 A mouse model study showed that acute‐phase mild hypertension induced by pharmaceuticals could lead to increased collateral cerebral blood flow, thereby improving oxygen metabolism in the penumbra. 104 Therefore, it may not be acute‐phase BP control but daily BP control that could contribute to better collateral status. Further high‐quality clinical studies are needed to clarify the association between admission BP and collateral status to guide management of BP during the acute phase of stroke. Both diabetes and high admission glucose levels were significantly associated with poor collateral status. It was reported that acutely evaluated glucose could impair nitrogen oxide availability, increase the production of endothelial‐derived vasoconstrictor prostanoids, and attenuate endothelium‐dependent vasodilation, which could lead to unfavorable collateral status. 105 , 106 Recent studies also found that chronically elevated blood glucose could impair leptomeningeal arteriogenesis in a type 2 diabetes transgenic mouse model. 107 , 108 These results imply that both acute‐phase and daily glucose control are needed to maintain the normal function of collateral flow. Our study also indicates that collateral status correlated with the underlying stroke mechanism. Cardioembolic stroke usually has an abrupt onset without chronic cerebral large vessel hypoperfusion, which may be associated with rare collateral artery formation. It was observed that internal carotid artery occlusion in the anterior circulation is associated with poor collateral status compared with more distal occlusion. The more proximal occlusion seems to increase the volume of the ischemic brain parenchyma and reduce the cerebral blood flow in the ischemic core, leading to depletion of the collateral vessels. The circle of Willis is a ring‐like intracranial arterial network that maintains sufficient blood supply to the brain, especially in conditions where there is occlusion in major arteries. However, we did not find that the integrity of the circle of Willis could significantly affect collateral status, which may be explained by the relatively small sample size. Understanding of the molecular pathogenesis of collateral status is vital to identify new therapeutic strategies either preventively before stroke onset or in the acute phase of stroke. Although recent preclinical studies have explored several molecular mechanisms of collateral circulation status, 62 , 109 , 110 the results of clinical studies in patients with acute ischemic stroke have been disappointing. 8 As several determinants of collateral status reported in our study, including older age, male sex, hypertension, and diabetes could be partly explained by endothelial dysfunction. 111 Ameliorating endothelial dysfunction might be a promising therapeutic target. Limitations Study heterogeneity is a main limitation of our results. This is because we included studies with different countries and races. In addition, both digital subtraction angiography and computer tomography angiography were used to evaluate collateral status, which increase the heterogeneity of the results. Thus, we performed subgroup analyses to identify the sources of heterogeneity. Moreover, some candidate factors associated with collateral status were reported in no more than 2 original studies, and we did not assess these candidate factors. In addition, due to the different anatomical structure between anterior circulation and posterior circulation in collateral status, we did not study the factors associated with collateral status in patients with posterior circulation stroke. In summary, because all studies included in this systematic review were observational studies and the overall quality of the evidence evaluated by the Grading of Recommendations Assessment, Development, and Evaluation assessment was rated as “very low,” further high‐quality therapeutic studies that focus on controlling risk factors are needed to support our findings. CONCLUSIONS Our systematic review and meta‐analysis revealed several factors could be associated with the degree of collateral status. Risk factors for atherosclerosis, including hypertension and diabetes could be modifiable determinants of collateral status. Although baseline characteristics, such as age, could be nonmodifiable risk factors, these could still be useful in the selection of ideal patients for collateral augmentation treatment. Further high‐quality therapeutic studies that focus on controlling risk factors are needed to support our findings. Sources of Funding This work was supported by the National Key R&D Program of Sichuan (2021YFG0316) to X.X., Start‐up Funds of the First Affiliated Hospital of Chengdu Medical College (CYFY‐GQ67) to K.L. Disclosures None. Supporting information Data S1 Tables S1–S3 Figures S1–S3 This article was sent to Jose R. Romero, MD, Associate Editor, for review by expert referees, editorial decision, and final disposition. Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.124.034170 For Sources of Funding and Disclosures, see page 6. Contributor Information Xun Xia, Email: xun.xia@aliyun.com. Lan Wen, Email: l_wen1225@163.com. References Associated Data Supplementary Materials Data S1 Tables S1–S3 Figures S1–S3