Patients with their first AIS (to exclude the effects of previous AIS events on circulating sCD163 concentrations) who visited the Department of Neurology, Chongqing General Hospital, University of Chinese Academy of Sciences during January 1, 2019, and January 31, 2020, were screened. The inclusion criteria included: (1) patients with newly onset AIS; (2) who visited the hospital within 24 h after symptom onset; (3) who are willing to participate. Seventy-eight age and sex-matched healthy subjects were recruited as controls from the healthy examination center of the same hospital. Subjects with diseases that might influence the circulating sCD163 levels were excluded from participation. Therefore, subjects were excluded if they have: (1) co-existing infections; (2) any types of tumors; (3) any types of autoimmune diseases; (4) other conditions that may influence the blood sCD163 levels, such as cirrhosis and severe inflammatory diseases; (5) declined to participate in this study. The patients were screened for eligibility for participation right after admission. However, although diabetes mellitus may contribute to the alteration of sCD163 (12), patients with diabetes mellitus were not excluded as it is a significant risk factor of AIS. Finally, 300 patients and 78 healthy controls were recruited in this study after excluding the subjects who failed for inclusion. Written informed consents for participation in this study and blood sampling were obtained from the subjects or their legal relatives. This study conformed with the principles of the Declaration of Helsinki and was approved by the investigational review board of the Chongqing General Hospital, University of Chinese Academy of Sciences.

At baseline (time of patient admission), the demographic information such as age, sex, and body mass index (BMI), pre-stroke medical history including oral antiplatelet or anticoagulants drug use, comorbidities including hypertension, diabetes mellitus, hypercholesteremia, and atrial fibrillation were collected and assessed right after admission. AIS was diagnosed according to the WHO Multinational Monitoring of Trends and Determinants in Cardiovascular Disease (WHO-MONICA) criteria with verification by MRI performed within 12 h after admission. The neurological deficits were examined with the National Institutes of Health Stroke Scale (NIHSS) upon admission (13), performed by a certified stroke neurologist. The AIS subtype was determined with the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria (14).

The patients were followed up for 1 month since admission for the observation of functional outcomes. The primary endpoint was neurological functions 1 month since admission. The functional outcomes were assessed using NIHSS and the modified Rankin Scale (mRS) (15), which were blinded to the plasma sCD163 concentrations.

Blood sampling was conducted right after patient admission before they receive any medical interventions. The blood was centrifuged right after collection at 20°C, spun at 2,000 g for 10 min, and stored at −80°C for biochemical analysis. The plasma sCD163 levels were determined using a human CD163 ELISA kit (Abcam, USA) according to the manufactural instructions. To preserve the linearity of the assays, samples containing high concentrations of sCD163 were diluted with an appropriate amount of calibrator diluent. The minimum detectable dose of sCD163 was 1.377 ng/ml, which was significantly lower than the sCD163 concentrations of the subjects in this study. Each test was conducted in duplicates and the means were used for statistical analysis.

If continuous variables were normally distributed, an independent t-test was used, and if not, a Mann-Whitney U-test was used. Two-sample tests of proportions were used to compare the proportions for the categorical variables. The comparisons of means among groups were conducted using one-way ANOVA and the comparisons of the rate of NIHSS change during the follow-up between groups were conducted using two-way ANOVA. Spearman correlation analyses were conducted to assess the association between plasma sCD163 concentrations and NIHSS or infarction volume. A linear regression model was utilized to investigate the association between the plasma sCD163 levels and functional improvement (as indicated by the change of NIHSS during follow-up). A logistical regression model was utilized to evaluate the risk factors of poor functional outcome (as indicated by mRS≥4) during follow-up. We first fitted univariate models with a single candidate variable at one time. The potential risk factors as determined by p < 0.05 were included in the final multivariate regression model. Statistical analyses were conducted using SPSS statistical package version 24 (IBM SPSS Statistics for Windows, Armonk, NY, USA).

Three-hundred-nine patients with AIS and 99 healthy controls were screened for eligibility for participation in this study. Two-hundred-nine patients failed screening due to the following reasons: 176 patients declined to participate, 2 patients deceased during hospitalization, 11 patients had co-existing infections, 13 patients had previously diagnosed tumors, and 7 patients had autoimmune diseases. Twenty-one healthy controls declined to participate in this study. Finally, 300 AIS patients and 78 healthy controls participated in this study (Figure 1).

The patients were further divided into the high and low plasma sCD163 groups, respectively, according to their plasma sCD163 concentrations. The patients had significantly higher median BMI than the healthy controls. No significant differences in the mean age, percentages of males, frequencies of subjects with a history of smoking, antiplatelet drug use, antithrombotic drug use, a family history of stroke, hypertension, diabetes mellitus, hypercholesteremia, and atrial fibrillation between the patients and controls were observed. The high sCD163 group had a significantly higher proportion of males, a higher median diffusion-weighted imaging (DWI) hyperintensity volume, higher NIHSS scores at baseline, and higher frequencies of lacunar-type stroke than the low sCD163 group. However, there was no difference in the mean age, median BMI, frequency of smoking, antiplatelet drug use, antithrombotic drug use, family history of stroke, comorbidities including hypertension, diabetes mellitus, hypercholesteremia, and atrial fibrillation between the two groups. Besides, the median DWI hyperintensity volume, the median NIHSS and mRS scores at follow-up, frequencies of atherothrombotic, cardioembolic, and unknown type stroke, frequencies of hemorrhagic transformation, and recurrent AIS during follow-up were not significantly different between the high and low sCD163 groups (Table 1).

We first compared the sCD163 concentrations between the patients with AIS and the control. We found that patients with AIS had significantly higher plasma sCD163 concentrations than the control (mean ± SD: 618.1 ± 292.0 ng/ml vs. 408.8 ± 157.5 ng/ml, p < 0.001) (Figure 2A). The high sCD163 group had better functional improvement as indicated by the change of NIHSS during follow-up than the low sCD163 group (p < 0.001) (Figure 2B). The plasma sCD163 concentrations were positively associated with the NIHSS at baseline (γ = 0.609, p < 0.001) and the DWI hyperintensity volume (γ = 0.509, p < 0.001) (Figures 2C,D). However, plasma sCD163 concentrations were not significantly associated with NIHSS at follow-up (γ = 0.049, p = 0.401) (Figure 2E). Patients with an mRS = 3 had significantly higher sCD163 concentrations than patients with an mRS = 1 (p < 0.001), 2 (p < 0.001), 4 (p < 0.001), and 5 (p < 0.001) (Figure 2F). Plasma sCD163 concentrations were positively associated with peripheral monocyte count, but no significance had been achieved (γ = 0.108, p = 0.062) (Figure 2G).

We utilized a linear regression model to investigate potential factors that were associated with plasma sCD163 levels. We found that hypercholesteremia and DWI hyperintensity volume were significantly associated with plasma sCD163 levels. These two factors along with diabetes mellitus, which is well-validated to be associated with sCD163 levels, were excluded from subsequent analysis (Table 2).

We first utilized a linear regression model to investigate the association between plasma sCD163 levels and functional improvement during follow-up as indicated by the decrease of NIHSS. Factors that are potentially associated with sCD163 levels were excluded from the model, including hypercholesteremia, DWI hyperintensity volume, and diabetes mellitus. Male sex and higher sCD163 concentrations were found to be positively associated with functional improvement during follow-up (Table 3).

We next utilized a logistical regression model to investigate the association between plasma sCD163 concentrations and poor functional outcome during follow-up as indicated by mRS ≥ 4. In the univariate analyses, stroke type was found to be associated with poor functional outcomes during follow-up. Higher sCD163 concentrations were found to be protective factors of poor functional outcome during follow-up, and these associations remained significant in the multivariate analyses (Table 4).