High levels of immuno-inflammatory markers predicts unfavorable short-term outcomes in patients with acute ischemic stroke

Atherosclerosis is a chronic inflammatory process that occurs in the arterial wall. This immuno-inflammatory process plays a role throughout all stages of stroke. Neutrophils, lymphocytes, and platelets are crucial blood cells for innate and adaptive immunity. This study investigated the associations of four types of immuno-inflammatory markers, namely the systemic immune-inflammation index (SII), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and neutrophil count (NC), with clinical outcomes in patients with acute ischemic stroke. Methods In this retrospective study, we enrolled 2903 inpatients with acute ischemic stroke from May 2010 to May 2019. Data included risk factors, laboratory parameters, and clinical features during hospitalization. The National Institutes of Health Stroke Scale (NIHSS), and modified Rankin Scale (mRS) were used to assess stroke severity and outcomes.


KEYWORDS
acute ischemic stroke, neutrophil count, neutrophil-to-lymphocyte ratio, platelet-tolymphocyte ratio, systemic immune-inflammation index, unfavorable outcome Introduction Stroke was the fourth leading cause of death in Taiwan from 2000 to 2020, and it is the leading cause of prolonged disability among older adults. Traditional risk factors for vascular diseases, such as old age, hypertension, diabetes mellitus, and heart disease, are prominent comorbidities of stroke. Most previous studies have emphasized the correlation of these comorbidities with stroke and clinical outcomes. The clinical feature of initial stroke severity has been reported to be a strong predictor of functional outcomes [1,2]. Laboratory parameters during acute stroke, such as hemoglobin level [3], blood urea nitrogen-to-creatinine ratio [4], and troponin I level [5], also provide valuable information when investigating clinical outcomes after stroke. Atherosclerosis is a chronic inflammatory process that occurs in the arterial wall [6]. Immunity and inflammation have been recognized as crucial elements of the pathobiology of stroke. Immuno-inflammatory processes play roles throughout all stages of acute stroke, including initial arterial occlusion, brain parenchymal damage, subsequent tissue repair, and infectious complications [7].
Innate and adaptive systems are two main types of immune systems. Innate immunity refers to the immune responses present at birth, and this provides first rapid defense against invasion. Innate immunity is mainly provided by neutrophils, monocytes, macrophages, natural killer cells, and complement systems [8]. Adaptive immunity, also known as acquired immunity, is provided by lymphocytes, which deliver antigen-dependent and antigen-specific responses to invasion. Recent studies have suggested that platelets not only mediate hemostasis and thrombosis but also participate in immuno-inflammatory responses by promoting innate effector cell functions and enhancing adaptive immune responses [9,10]. To clarify the associations between immunoinflammatory responses and outcomes of acute stroke, several clinical studies have investigated neutrophils, lymphocytes, and platelets. Detection of a single cellular line may be insufficient to recognize the status and complexity of the immune system. Moreover, a single blood cell test can be influenced by conditions such as overhydration, dehydration, and the treatment of blood specimens [11]. Ratios of cell measurements, such as the neutrophil-to-lymphocyte ratio (NLR) and platelet-tolymphocyte ratio (PLR), are considered suitable for reflecting the balance between innate and adaptive immunity [12]. Higher  However, the correlation of SII with stroke outcomes has rarely been explored. In the present study, we aimed to investigate the association of four immuno-inflammatory markers, namely SII, NLR, PLR, and NC, with clinical outcomes in patients with acute ischemic stroke. We also compared the predictive performance of these four markers.

Study Population and Data Collection
The stroke registry database was retrospectively reviewed to identify patients who were treated for stroke in a neurological ward from May 2010 to May 2019. Inclusion criteria were 1) diagnosis of acute ischemic stroke confirmed by clinical presentation and 2) proof of an ischemic lesion or absence of a corresponding intracranial lesion other than infarction according to brain computed tomography or magnetic resonance imaging. Sex, age, history of hypertension, diabetes mellitus, hyperlipidemia, heart disease, prior stroke, smoking status, alcohol consumption, cancer, uremia, and length of stay (LOS) in hospital were recorded for analysis. Laboratory data obtained on arrival in the emergency department included full blood count with white blood cell differentials as well as platelet, glucose, creatinine, and troponin I levels. Abnormal elevation of troponin I was defined as a blood troponin I level > 0.01 μg/L; troponin I level was registered as 0.01 μg/L if the patient had a value of ≤ 0.01 μg/L. Fasting cholesterol and triglyceride were recorded during the morning after admission to the ward.

Stroke Severity and Clinical Features
Stroke severity was assessed on admission according to the National Institutes of Health Stroke Scale (NIHSS). We classified the etiology of ischemic stroke according to the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) categories, namely large-artery atherosclerosis, small-vessel occlusion, cardioembolism, other determined etiology, and undetermined etiology [20]. Urinary tract infection, pneumonia, gastrointestinal bleeding, and seizure were registered as in-hospital stroke complications.
Functional outcomes were evaluated using the NIHSS, the Barthel index and the modified Rankin Scale (mRS) at discharge. An mRS score > 2 was considered to indicate an unfavorable outcome.

Definition of Immuno-inflammatory Markers
Four immuno-inflammatory markers, namely SII, NLR, PLR, and NC, were obtained for analysis. The NLR and PLR were calculated as the ratio of neutrophil count to lymphocyte count and the ratio of platelet count to lymphocyte count, respectively. SII was calculated as the platelet count multiplied by the NLR. For patients admitted to the ward with transient ischemic attack (TIA) during the same period, these immuno-inflammatory markers were also collected for comparison.

Statistical Analysis
Continuous variables are presented as mean ± standard deviation.

Results
During the study period, 2903 patients with acute ischemic stroke and 457 patients with TIA were enrolled. Of the 2903 patients with acute ischemic stroke, 2036 patients (70%) had valid data concerning troponin I levels because troponin I was not routinely measured in the emergency department for patients with acute stroke during that period. The average age of patients with acute ischemic infarct and patients with TIA were 71.0 ± 13.5 and 70.5 ± 13.2 years, respectively. Table 1 shows the clinical characteristics of the 2903 patients with acute ischemic stroke and the sex differences for these characteristics. The average values of SII, NLR, PLR, and NC were 892 ± 1354, 4.2 ± 56, 145 ± 119, and 5.6 ± 2.6 × 10 3 /mL, respectively. Female patients tended to be older than their male counterparts; they had higher platelet count, SII, PLR, cholesterol, LOS in hospital, NIHSS on admission, and mRS at discharge; and they had higher rates of heart disease, history of cancer, and in-hospital complications. Male patients had higher levels of hemoglobin, creatinine, and triglyceride than their female counterparts, as well as higher rates of prior stroke, smoking, and alcohol consumption. Table 2 shows correlations of the four immuno-inflammatory markers with the measured variables. All of the immuno-inflammatory markers exhibited similar linear correlations with the measured variables, including positive correlations with age, glucose level, creatinine level, LOS in hospital, NIHSS on admission, NIHSS at discharge, and mRS at discharge; negative correlations were observed with hemoglobin level, triglyceride level (except for NC), and Barthel index at discharge. By using ROC curve analysis, we identified cutoff points for the immuno-inflammatory markers to indicate unfavorable outcomes (i.e., mRS >2). The cutoff points for SII, NLR, PLR, and NC were 724, 3.5, 143, and 6 × 10 3 /mL, respectively. Patients with SII, NLR, PLR, and NC higher than cutoff values also had higher rates of history of cancer (except for SII and NC), uremia (except for NC), elevated troponin I level, in-hospital complications, and unfavorable outcomes ( Table 2).
We compared the levels of the four immuno-inflammatory markers in the five groups categorized according to the TOAST classifications and patients with TIA; this comparison revealed the same results (Table 3). Patients with cardioembolism tended to be older and the highest initial NIHSS scores, whereas patients with small-vessel disease had the lowest initial NIHSS scores. No difference was observed in the levels of the four immuno-inflammatory markers between patients with TIA and those with small-vessel disease, or between those with large-artery atherosclerosis and those with cardioembolism. Levels of the four markers were significantly higher in patients with large-artery atherosclerosis, cardioembolism, and other determined etiology compared with other patients (p ≤ 0.002). The highest levels of these markers were observed in patients with other determined etiology.
Univariate analyses of continuous variables revealed that older age; higher levels of white blood cell counts, SII, NLR, PLR, NC, and creatinine; longer LOS in hospital; higher NIHSS on admission; and lower levels of hemoglobin, cholesterol, and triglyceride were significantly associated with unfavorable outcomes (Table 4). Univariate analyses of dichotomous variables revealed that female sex, hypertension, diabetes mellitus, heart disease, prior stroke, history of cancer, uremia, elevated troponin I level, and in-hospital complications were significant positive predictors of unfavorable outcomes. However, hyperlipidemia, smoking status, and alcohol consumption were positive predictors of favorable outcomes. Table 5 presents the regression analysis for the effect of the main significant factors in Table 4  The C-statistics of the regression models for the detection of unfavorable outcomes are shown in Table 6 for each predictor. We established a basic model comprising the seven significant predictors of admission NIHSS ≥ 5, age > 75 years, diabetes mellitus, elevated troponin I level, female sex, heart disease, and prior stroke. The C-statistic for this basic model was 0.860 ( Figure 1). The addition of SII to the basic model resulted in a significant improvement in the C-statistic from 0.860 to 0.864 (p = 0.038; Table 6). The addition of NLR or PLR to the basic model also resulted in significant improvements of the C-statistic from 0.860 to 0.863 (p = 0.042 and 0.048 for NLR and PLR, respectively). Including NC in the basic model improved the C-statistic from 0.860 to 0.861, but the difference was not significant. The simultaneous addition of all four immuno-inflammatory markers to the basic model resulted in a significant improvement of the C-statistic to 0.864 (p = 0.032); the addition of all markers did not result in a stronger predictive performance than addition of SII alone.

Discussion
Sex differences in the risk factor distribution, severity, and outcomes of ischemic stroke are multifactorial and related to genetics, environmental factors, and social influences [21][22][23]. Previous studies have revealed that older women with higher stroke severity at stroke onset have higher platelet counts, higher prevalence of cardioembolism, and more unfavorable clinical outcomes.
Similar results were observed in the present study. Furthermore, we found that women had higher SII and PLR, as well as higher rates of cancer history, elevated troponin I level, and in-hospital complications. No differences in NLR and NC were observed between male and female patients, possibly due to absence of platelet count in these markers. Recent studies have revealed that risk of stroke not only increases after a new cancer diagnosis but also increases with time in almost all cancer survivors [24,25]. Cancer and related therapies may cause coagulopathies, such as nonbacterial thrombotic endocarditis, alterations in platelet and endothelial function, and radiationinduced atherosclerosis. Elevated troponin I level during acute stroke is a strong independent predictor for both unfavorable outcomes and in-hospital mortality. The mechanisms of elevated troponin I level during acute stroke include ischemic myocardial injury, neurogenic heart syndrome through increased sympathetic activity causing cardiomyopathy, and other systemic conditions such as infection, sepsis, renal failure, and pulmonary embolism [5].
Atherosclerosis is the primary underlying pathological process in coronary and cerebral arterial diseases; it is considered as a chronic inflammation that causes large and medium arterial thromboses [26]. The innate and adaptive immune mechanisms are both involved in the prothrombotic progression of atherosclerotic change. When acute ischemic stroke occurs during arterial occlusion, the inflammatory response following the release of danger signals from damaged brain tissue leads to an activation of immune system. Innate immunity, including neutrophils, monocytes, macrophages, platelet, and dendritic cells, is rapidly activated with the production of various cytokines. This is followed by activation of the adaptive immunity, namely lymphocytes, which exerts an immunosuppressive effect that promotes intercurrent infections (i.e., stroke-induced immunodepression) [7]. These immunological changes may last for weeks and may increase the risk of respiratory or urinary tract infections, particular among patients with severe stroke, thus affecting clinical outcomes [27]. Neutrophils, which are secretory and phagocytic cells, migrate to the intraparenchymal perivascular areas within several hours after cerebral ischemia and participate in the early destruction of the blood-brain barrier [28]. Higher NC indicates a larger area of ischemia and more severe brain damage. Lymphocytes, which mainly comprise humoral immune response B cells and cellular immunity T cells, accumulate in the brain 3-6 days after stroke and are considered as having a regulatory function by inducing neuroprotection. Persistent lymphopenia after stroke, caused by the redistribution of lymphocytes to the lymphatic organs and increased catecholamine and cortisol levels, indicates prolonged brain damage with a higher stress response, and this is associated with unfavorable long-term prognosis [29]. In addition to promoting the progression of atherosclerosis, platelets release mediators to boost inflammation after stroke and result in the release of neutrophils and lymphocytes into the vessel wall. For patients with cancer, neutrophils and platelets have also been observed to promote cancer cell proliferation, invasion, immune evasion, and metastasis through multiple mechanisms. Therefore, elevated levels of inflammatory markers are considered to indicate a substantial tumor burden and an ongoing chronic inflammatory process [30].
In the present study we found that all four immuno-inflammatory markers were positively correlated with age, glucose level, creatinine level, NIHSS on admission, LOS in hospital, and mRS at discharge.
Patients whose immuno-inflammatory markers were higher than the cutoff values for unfavorable outcomes also exhibited higher rates of uremia, elevated troponin I level, and in-hospital complications. Higher NLR and PLR have been reported in patients with type 2 diabetes mellitus and hyperglycemia, respectively [31,32]. Although previous studies have suggested that these immunoinflammatory markers were increased among patients with various cancers, we did not identify any differences among stroke patients with and without a history of cancer, possibly because these cancers were inactive or cured. Cholesterol level had an inverse correlation with age and immuno-  6-1.8). The predictive performance for unfavorable outcomes was similar when using SII, NLR, and PLR, whereas the performance of NC was slightly weaker. Because the four markers were derived from white blood cell counts with or without platelet counts, which are essential laboratory data during acute stroke and common routine examinations, we can choose one as a reference marker for the prediction of unfavorable outcomes.
SII > 724 is the most appropriate marker because this provided the optimal predictive performance of 0.864 when combined with the other seven predictors.
This study had several limitations. First, this was a retrospective study. We did not have sufficient sequential data during hospitalization for a dynamic comparison of immuno-inflammatory markers. A dynamic increase in NLR has been reported to predict 3-month mortality or major disability among patients receiving intravenous thrombolytic treatment [33]. Second, we did not investigate the association between infarct volume and the immuno-inflammatory markers. However, the TOAST classifications may partly reflect the infarct size. Third, because we did not perform a follow-up study after discharge, only short-term outcomes at discharge were available. A prospective study with serial immuno-inflammatory markers and long-term outcomes may provide more prognostic relevance for acute ischemic stroke. Notwithstanding these limitations, the results extend the current understanding of the implications of immuno-inflammatory markers among patients with acute ischemic stroke.

Conclusion
Initial stroke severity (NIHSS on admission ≥ 5) and age (age > 75 years) are the two most significant predictors of unfavorable outcome among patients with acute ischemic stroke. Immuno-inflammatory markers including SII, NLR, PLR, and NC provide improved prediction of stroke outcomes compared with conventional risk factors and laboratory parameters. SII > 724 is the most appropriate marker

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.      This is a list of supplementary files associated with this preprint. Click to download.