Red Blood Cell Distribution Width Is Associated with Collateral Flow and Final Infarct Volume in Acute Stroke with Large Artery Atherosclerosis

Red blood cell distribution width (RDW), routinely assessed as part of the complete blood count, is an acknowledged index of poor outcomes in people with cardiovascular diseases and stroke, independent of anemia and inflammatory status.[1] [2] [3] However, few studies have explored mechanisms underlying associations between higher RDW and worse clinical outcomes in stroke patients. Hypotheses include that RDW represents an inflammatory or oxidative marker and that RDW correlates with baseline systemic diseases.[2] [3]

In patients with non-ST elevation myocardial infarction, high RDW correlates with insufficient collateral supply and poor outcomes.[4] Since large artery atherosclerosis (LAA) related stroke shares pathogenesis similar to acute coronary artery syndrome, we hypothesized that worse clinical outcomes with higher baseline RDW for LAA stroke could at least be partly related to collateral flow. Our hypothesis is that for LAA stroke, higher baseline RDW is associated with poorer collateral flow and increased final infarct volume (FIV), therefore leading to worse clinical outcomes.

We evaluated our hypothesis by the assessment of consecutive acute ischemic stroke patients presenting within 6 hours of symptom onset to Huashan Hospital, Fudan University, China, and Shanghai East Hospital, Tongji University, China, from November 2011 to December 2018. Patients were included in this study if they (1) underwent complete baseline multimodal computed tomography (CT) imaging including noncontrast CT (NCCT), CT angiography, and CT perfusion, (2) had complete baseline clinical profiles and follow-up data, and (3) their stroke subtype was identified as LAA according to the TOAST (Trial of Org 10172 in Acute Stroke Treatment) classification.[5] This study was approved by the ethics committees of both participating hospitals, and written informed consent was obtained from each patient.

Baseline perfusion images were postprocessed using commercial software (MIStar, Apollo Medical Imaging Technology). Previously validated thresholds were applied to measure the volume of acute hypoperfused lesion (relative delay time [DT] > 3s), severely hypoperfused lesion (DT > 6s), and baseline infarct core volume (relative cerebral blood flow [CBF] < 30%) with single value deconvolution with delay and dispersion correction.[6] [7]

Collateral flow was evaluated using the volume ratio of DT > 3s perfusion lesion with tissue with severely delayed contrast transit (DT > 3s/DT > 6s).[8] FIV was also calculated on follow-up images (magnetic resonance imaging [MRI]/CT) on MIStar by semiautomatically drawing regions of interest. Infarct volume growth was determined as FIV minus the volume of estimated baseline infarct core.

The RDW was measured at patients' arrival at the emergency room using Sysmex XN 2000 (Sysmex Corporation) in Huashan Hospital and Sysmex XS 1000i (Sysmex Corporation) in Shanghai East Hospital, with ethylenediaminetetraacetic acid (EDTA) blood samples. Though no direct comparison has been made between these two instruments, both have been shown to have good consistency with Sysmex XE 2100.[9] [10]

The association between RDW, collateral flow, FIV and infarct volume growth was evaluated using univariate and multivariate generalized linear models, with collateral flow (DT > 3s/DT > 6s) and FIV being log-transformed. Variables with p < 0.05 in univariate analyses or with clinical significance were included in multivariate models. Multicollinearity for RDW was absent in all models, as proven by the assessment of variance inflation factors.[11]

A total of 112 patients with LAA stroke were included in this study, of whom 82 (73.2%) patients were male; the median age was 65 (interquartile range [IQR] = 59.0–74) years, with a median baseline NIHSS (National Institutes of Health Stroke Scale) of 8.0 (4.0–13.8). The median value of DT > 3s/DT > 6s volume ratio was 4.6 (IQR = 2.7–9.6). The mean value of hemoglobin was 143.1 g/L (standard deviation = 17.8 g/L). The median volume of final infarct area was 12.6 mL (IQR = 5.6–33.6 mL). The median value of RDW was 12.8% (IQR = 12.4–13.2%). A total of 104 patients underwent follow-up MRI/NCCT scans, and there were no discernible differences in the clinical and laboratory test profiles between patients with follow-up images and the total study population. A comparison between different RDW level groups (divided by RDW median value of 12.8%) is listed in [Table 1]. Although no statistical significance was shown, patients with elevated RDW demonstrated larger infarct core, DT > 3s lesion, DT > 6s lesion, and FIV, as well as a poorer collateral flow.

Table 1 Comparison of baseline, follow-up, and laboratory test profiles for patients with different RDW levels[a]

	RDW < 12.8% (n = 56)	RDW > 12.8% (n = 56)	p-Value
Demographic data and outcomes
Male sex	44 (78.6%)	38 (67.9%)	0.20
Age, median (IQR), years	62.0 (59.0–70.0)	68.5 (59.3–75.8)	0.07
Baseline NIHSS, median (IQR)	7.0 (4.0–13.8)	9.0 (4.0–13.8)	0.67
Medical history
Smoking	31 (55.4%)	25 (44.7%)	0.26
Hypertension	40 (71.4%)	44 (78.6%)	0.38
Dyslipidemia	14 (25.0%)	10 (17.9%)	0.38
Diabetes	19 (33.9%)	19 (33.9%)	1.00
Stroke history	13 (23.2%)	15 (26.8%)	0.66
Congestive heart failure	2 (3.6%)	2 (3.6%)	1.00
Coronary artery atherosclerosis	2 (3.6%)	5 (8.9%)	0.44
Anemia[b]	1 (1.8%)	5 (8.9%)	0.21
Taking antiplatelet before admission	11 (19.6%)	10 (17.9%)	0.81
Taking anticoagulant before admission	1 (1.8%)	2 (3.6%)	1.00
Taking statin before admission	9 (16.1%)	3 (5.4%)	0.07
Reperfusion therapy	47 (83.9%)	43 (76.8%)	0.34
IV rtPA	43 (76.8%)	34 (60.7%)	0.07
Endovascular treatment	11 (19.6%)	11 (19.6%)	1.00
3m mRS 0–1	23 (41.1%)	23 (41.1%)	1.00
3m mRS 0–2	33 (58.9%)	33 (58.9%)	1.00
Hemorrhagic transformation	10 (17.9%)	5 (8.9%)	0.17
PH2	4 (7.1%)	1 (1.8%)	0.37
CTP data
Core volume, median (IQR), mL	5.9 (1.1–23.4)	8.7 (2.8–17.7)	0.43
DT > 3s lesion volume, median (IQR), mL	44.1 (19.9–120.6)	83.0 (33.3–143.2)	0.19
DT > 6s lesion volume, median (IQR), mL	7.4 (2.1–35.7)	13.7 (5.0–44.2)	0.08
DT > 3s / DT > 6s, median (IQR)	4.77(3.0–12.7)	4.1 (2.5–7.7)	0.16
Final infarct volume, median (IQR), mL	10.3 (3.3–32.8)	14.1 (5.9–38.6)	0.26
Infarct volume growth, median (IQR), mL	4.0(–2.5 to 13.3)	6.3 (–0.2 to 18.9)	0.17
Laboratory tests
Baseline glucose, median (IQR), mmol/L	6.7 (5.7–9.5)	6.6 (5.7–8.2)	0.20
White blood cell count, mean (SD), ×109/L	7.6 (6.6–9.6)	8.1 (6.4–10.2)	0.65
Red blood cell count, mean (SD), ×1012/L	4.7 (0.5)	4.6 (0.5)	0.19
Platelet count, mean (SD), ×109/L	220.5 (48.1)	201.2 (76.6)	0.11
Hemoglobin, mean (SD), g/L	146.5 (15.9)	139.6 (19.1)	0.04
Hematocrit, mean (SD), %	42.8 (4.4)	41.7 (4.6)	0.20
Monocyte, median (IQR), %	6.1 (4.7–7.8)	6.0 (5.0–7.0)	0.80
Lymphocyte, median (IQR), %	22.0 (14.7–27.6)	18.4 (14.3–25.1)	0.26
Neutrophil, median (IQR), %	68.2 (61.8–78.2)	72.7 (63.6–78.5)	0.25
Creatinine, median (IQR), umol/L	69.0 (59.8–78.0)	72.0 (59.5–87.8)	0.44
Albumin, median (IQR), g/L	43.5 (41.0–46.0)	42.0 (39.0–43.0)	0.003
Fibrinogen, median (IQR), g/L	2.6 (2.2–3.0)	2.7 (2.3–3.4)	0.11

Abbreviations: CTP, computed tomography perfusion; DT, delay time; IQR, interquartile range; IV rtPA, intravenous recombinant tissue plasminogen activator; mRS, modified Rankin scale; NIHSS, National Institutes of Health Stroke Scale; PH2, parenchymal hematoma type 2; RDW, red blood cell distribution width; SD, standard deviation.

^a Data are presented as number (percentage) of patients unless otherwise indicated.

^b Anemia was identified as baseline hemoglobin < 120g/L (male) or 110g/L (female).

The scatter plots of RDW, log-transformed DT > 3s/DT > 6s ratio, log-transformed FIV, and infarct volume growth are shown in [Fig. 1]. Higher RDW is associated with smaller DT > 3s/DT > 6s ratio, indicating poorer collateral flow, and the association remains significant after multivariate adjustment (unadjusted coefficient: –0.46, 95% CI: –0.55 to –0.38, p < 0.001; multivariate adjusted coefficient: –0.57, 95% CI: –0.67 to –0.47, p < 0.001, [Table 2]). For 104 patients with follow-up images, every 1% increase of RDW was associated with increasing FIV (unadjusted coefficient: 0.42, 95% CI: 0.38–0.46, p < 0.001; multivariate adjusted coefficient: 0.06, 95% CI: 0.01–0.11, p = 0.04, [Table 2]). Moreover, every 1% increase of RDW was associated with a 16.5-mL increase of infarct volume growth, whereas multivariate-adjusted analysis did not show statistical significance (unadjusted coefficient: 16.5, 95% CI: 0.4–32.5, p = 0.04; multivariate adjusted coefficient: 10.2 95% CI: –6.1 to 26.4, p = 0.22, [Table 2]).

Table 2 The association between RDW, collateral flow, final infarct volume, and infarct volume growth

	Unadjusted			Multivariate adjusted
RDW as a continuous variable (per 1% change)	Coefficient	95% CI	p-Value	Coefficient	95%CI	p-Value
DT > 3s/DT > 6s[a]	–0.46	–0.55 to –0.38	< 0.001	–0.57	–0.67 to –0.47	< 0.001
Final infarct volume[b]	0.42	0.38–0.46	< 0.001	0.06	0.01–0.11	0.04
Infarct volume growth[c]	16.5	0.4–32.5	0.04	10.2	–6.1 to 26.4	0.22

Abbreviations: CI, confidence interval; DT, delay time; RDW, red blood cell distribution width.

^a DT > 3s/DT > 6s was log-transformed. Multivariate model was adjusted for age, baseline glucose history of dyslipidemia, history of stroke, history of chronic heart failure, white blood cell count, red blood cell count, platelet count, hemoglobin, hematocrit, creatinine, albumin, fibrinogen, monocyte percentage, and lymphocyte percentage.

^b n = 104. Final infarct volume was log-transformed. Multivariate model was adjusted for age, baseline glucose, history of hypertension, history of chronic heart failure, red blood cell count, platelet, hemoglobin, hematocrit, creatinine, albumin, monocyte percentage, lymphocyte percentage, neutrocyte percentage, intravenous recombinant tissue plasminogen activator (rtPA), endovascular treatment, and magnetic resonance imaging (MRI) vs. noncontract computer tomography (NCCT).

^c n = 104. Multivariate model adjusted for history of chronic heart failure, albumin, intravenous rtPA, endovascular treatment, and MRI vs. NCCT.

Fig. 1 Scatter plot for the correlation between red blood cell distribution width (RDW, %), collateral flow, final infarct volume, and infarct volume growth. *Log-transformed. (A) Scatter plot for the correlation between RDW (%) and collateral flow (n = 112). (B) Scatter plot for the correlation between RDW (%) and final infarct volume (n = 104). (C) Scatter plot for the correlation between RDW (%) and infarct volume growth (n = 104).

The results of this study demonstrate that increased baseline RDW is associated with (1) poor collateral flow and (2) increased FIV in patients with LAA stroke. There are some proposed hypotheses on the relationship between high RDW and poor cerebral collaterals. First, studies have shown that high RDW is associated with elevated oxidative stress.[12] Elevated oxidative stress can induce red blood cell (RBC)/endothelial cell adherence and reduce RBC deformability, both of which play a role in the elevation of vessel resistance.[13] RBC deformability is one of the determinants of blood viscosity alterations, which was considered to be correlated with CBF maintenance failure under acute ischemic setting.[14] [15] Second, RDW is associated with deteriorated atherosclerosis through inflammatory mechanisms.[16] [17] Cerebral collateral flow is initiated promptly after occlusion, and the circle of Willis serves as the first pathway of collateral flow. Therefore, atherosclerotic arteries of the circle of Willis in patients with LAA directly confine the ability of rapid perfusion through this pathway. Third, increased RDW was related to increased B-type natriuretic peptide.[18] Research has also shown that middle cerebral artery stenosis or occlusion can promote distal vessel anastomoses of the secondary pathway (the leptomeningeal collaterals), providing retrograde flow.[19] It was also reported that natriuretic peptide inhibited angiogenesis, which might suppress the development of distal anastomoses.[20] Fourth, high RDW was also related to autonomic dysfunction, with reduced heart rate in patients with heart failure.[21] For patients with acute ischemic stroke, autoregulated blood pressure is of great importance to sustain collateral flow,[22] which might also be affected by autonomic dysfunction.

Apart from RDW and collateral flow, our study demonstrated that increased RDW was associated with enlarged FIV and increased infarct volume growth (though the latter was not statistically significant with multivariate adjustment). The relationship between FIV and RDW can be reasonably explained by the impaired collateral flow with increased RDW. Another possible explanation is that anisocytosis (measured by RDW) is associated with reduced oxygen-carrying capacity and reduced deformability,[3] [13] thus suggesting impaired oxygen supply in the brain tissue distal to the occluded artery. Furthermore, RDW might serve as a marker of the procoagulant status of RBCs, though its mechanism has not been clearly elucidated.[2] This procoagulant status may undermine the revascularization efficacy of intravenous thrombolysis and other reperfusion therapy, leading to infarct expansion.

There are some notable limitations of this study. First is the limited sample size. Second, not all patients underwent follow-up MRI/CT scans; however, no differences in the clinical and laboratory test profiles between patients with follow-up images and total study population was found. Lastly, this is a retrospective study, and we did not adjust for high-sensitivity C-reactive protein (hsCRP) in our multivariate models (RDW is known to be correlated with inflammatory status) since hsCRP was not included in the routinely performed laboratory test panel in the emergency room at Huashan Hospital or East Hospital. But we included some significant percentages of certain white blood cells in the multivariate model to avoid the influence of inflammation. Furthermore, it has been reported that the association strength between RDW and outcomes was not attenuated after hsCRP adjustment.[23]

Author Contributions

Xin Cheng, Gang Li, Mark Parsons, and Qiang Dong contributed to conception and design of the study. Yifeng Ling, Lumeng Yang, Feifeng Liu, and Wenjie Cao contributed to data collection and analysis. Lan Hong and Kun Fang contributed to drafting the text and preparing the tables and figures.

^* These two authors contributed equally.

• References

• 1 Salvagno GL, Sanchis-Gomar F, Picanza A, Lippi G. Red blood cell distribution width: a simple parameter with multiple clinical applications. Crit Rev Clin Lab Sci 2015; 52 (02) 86-105
• 2 Kim J, Kim YD, Song TJ. , et al. Red blood cell distribution width is associated with poor clinical outcome in acute cerebral infarction. Thromb Haemost 2012; 108 (02) 349-356
• 3 Turcato G, Cappellari M, Follador L. , et al. Red blood cell distribution width is an independent predictor of outcome in patients undergoing thrombolysis for ischemic stroke. Semin Thromb Hemost 2017; 43 (01) 30-35
• 4 Tanboga IH, Topcu S, Nacar T. , et al. Relation of coronary collateral circulation with red cell distribution width in patients with non-ST elevation myocardial infarction. Clin Appl Thromb Hemost 2014; 20 (04) 411-415
• 5 Adams Jr HP, Bendixen BH, Kappelle LJ. , et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 1993; 24 (01) 35-41
• 6 Bivard A, Levi C, Krishnamurthy V. , et al. Defining acute ischemic stroke tissue pathophysiology with whole brain CT perfusion. J Neuroradiol 2014; 41 (05) 307-315
• 7 Bivard A, Levi C, Lin L. , et al. Validating a predictive model of acute advanced imaging biomarkers in ischemic stroke. Stroke 2017; 48 (03) 645-650
• 8 Hong L, Cheng X, Lin L. , et al; INSPIRE Study Group. The blood pressure paradox in acute ischemic stroke. Ann Neurol 2019; 85 (03) 331-339
• 9 Seo JY, Lee ST, Kim SH. Performance evaluation of the new hematology analyzer Sysmex XN-series. Int J Lab Hematol 2015; 37 (02) 155-164
• 10 Ghys T, Malfait R, VAN den Bossche J. Performance evaluation of the Sysmex XS-1000i automated haematology analyser. Int J Lab Hematol 2009; 31 (05) 560-566
• 11 Chatterjee S, Hadi AS. Regression Analysis by Example. 4th ed. Hoboken, NJ: Wiley; 2006: 233-239
• 12 Semba RD, Patel KV, Ferrucci L. , et al. Serum antioxidants and inflammation predict red cell distribution width in older women: the Women's Health and Aging Study I. Clin Nutr 2010; 29 (05) 600-604
• 13 Kaul DK, Koshkaryev A, Artmann G, Barshtein G, Yedgar S. Additive effect of red blood cell rigidity and adherence to endothelial cells in inducing vascular resistance. Am J Physiol Heart Circ Physiol 2008; 295 (04) H1788-H1793
• 14 Grotta J, Ostrow P, Fraifeld E, Hartman D, Gary H. Fibrinogen, blood viscosity, and cerebral ischemia. Stroke 1985; 16 (02) 192-198
• 15 Allport LE, Parsons MW, Butcher KS. , et al. Elevated hematocrit is associated with reduced reperfusion and tissue survival in acute stroke. Neurology 2005; 65 (09) 1382-1387
• 16 Furer A, Finkelstein A, Halkin A. , et al. High red blood cell distribution width and preclinical carotid atherosclerosis. Biomarkers 2015; 20 (6-7): 376-381
• 17 Jia H, Li H, Zhang Y, Li C, Hu Y, Xia C. Association between red blood cell distribution width (RDW) and carotid artery atherosclerosis (CAS) in patients with primary ischemic stroke. Arch Gerontol Geriatr 2015; 61 (01) 72-75
• 18 Fukuta H, Ohte N, Mukai S. , et al. Elevated plasma levels of B-type natriuretic Peptide but not C-reactive protein are associated with higher red cell distribution width in patients with coronary artery disease. Int Heart J 2009; 50 (03) 301-312
• 19 Jashari F, Ibrahimi P, Nicoll R, Bajraktari G, Wester P, Henein MY. Coronary and carotid atherosclerosis: similarities and differences. Atherosclerosis 2013; 227 (02) 193-200
• 20 Pedram A, Razandi M, Levin ER. Natriuretic peptides suppress vascular endothelial cell growth factor signaling to angiogenesis. Endocrinology 2001; 142 (04) 1578-1586
• 21 Borné Y, Smith JG, Melander O, Engström G. Red cell distribution width in relation to incidence of coronary events and case fatality rates: a population-based cohort study. Heart 2014; 100 (14) 1119-1124
• 22 Jiang B, Churilov L, Kanesan L. , et al. Blood pressure may be associated with arterial collateralization in anterior circulation ischemic stroke before acute reperfusion therapy. J Stroke 2017; 19 (02) 222-228
• 23 Saliba W, Barnett-Griness O, Elias M, Rennert G. The association between red cell distribution width and stroke in patients with atrial fibrillation. Am J Med 2015; 128 (02) 192.e11-192.e18