Serum sCD163 as a potential biomarker for predicting poor functional outcome at 3 months in acute ischemic stroke without reperfusion: a prospective study.
Yan M, Shao L, Li Z, Liu C • Frontiers in immunology • 2026
Serum sCD163 within 24 hours of admission is a robust, independent biomarker for predicting 3-month poor functional outcome in AIS patients without reperfusion, and a prognostic model combining sCD163 with traditional clinical and imaging variables demonstrates excellent discrimination, calibration, and clinical utility in an independent external cohort.
Key Findings
Results
Poor functional outcome at 3 months occurred in approximately one-third of AIS patients without reperfusion therapy in both cohorts.
Poor outcome (mRS > 2 at 3 months) occurred in 34.1% (226/664) of the training cohort and 34.0% (102/300) of the external validation cohort.
Training cohort enrolled 664 first-ever AIS patients from 2021-2023; external validation cohort enrolled 300 patients from an independent center in 2024.
All patients did not receive intravenous thrombolysis or mechanical thrombectomy.
Prospective, dual-cohort study design was used.
Results
Serum sCD163 levels were significantly higher in patients with poor functional outcome compared to those with good outcome.
Mean sCD163 levels were 820 ± 110 ng/mL in poor-outcome patients versus 650 ± 80 ng/mL in good-outcome patients (P < 0.001).
Serum sCD163 was measured within 24 hours of admission by ELISA.
The optimal cutoff for sCD163 was determined to be 742.5 ng/mL.
sCD163 is described as a marker of monocyte/macrophage activation.
Results
High serum sCD163 (≥ 742.5 ng/mL) was an independent predictor of poor functional outcome after multivariable adjustment.
Adjusted OR = 7.52 (95% CI: 4.42–12.78, P < 0.001) for high sCD163 as a predictor of poor outcome.
LASSO variable selection was used prior to multivariable logistic regression to identify independent predictors.
Other independent predictors identified in the model included baseline NIHSS score, DWI infarct volume, and atrial fibrillation.
sCD163 remained independently predictive after adjustment for these clinical and imaging variables.
Results
The combined prognostic model incorporating sCD163 achieved excellent discrimination and calibration in both internal bootstrap validation and external validation.
Internal bootstrap-validated AUC was 0.877 (95% CI: 0.829–0.926).
External validation AUC was 0.867 (95% CI: 0.815–0.921).
Good calibration was demonstrated in both cohorts: Hosmer-Lemeshow P = 0.868 (training) and P = 0.680 (external validation).
Decision curve analysis showed superior net clinical benefit for the combined model.
Model performance was assessed by discrimination (AUC), calibration (Hosmer-Lemeshow test), and decision curve analysis (DCA).
Results
Adding sCD163 to a model with clinical predictors alone significantly improved prognostic performance.
The incremental improvement in AUC from adding sCD163 was ΔAUC = 0.118 (P < 0.001).
The combined model included sCD163 alongside traditional clinical and imaging variables.
Decision curve analysis confirmed superior net clinical benefit of the combined model over clinical predictors alone.
Background
Neuroinflammation via monocyte/macrophage activation, as reflected by sCD163, has been linked to stroke severity and was the mechanistic basis for this investigation.
sCD163 is characterized as a marker of monocyte/macrophage activation.
Neuroinflammation is described as playing a critical role in the prognosis of acute ischemic stroke.
Prior to this study, the independent prognostic value and clinical utility of sCD163 in AIS remained uncertain.
The authors note that further mechanistic and multi-center studies are needed to confirm generalizability and causality.
What This Means
This research suggests that a protein called sCD163, measured in the blood within 24 hours of a stroke, can help predict which patients are likely to have poor recovery three months later. sCD163 is released by immune cells called monocytes and macrophages when they become activated, and the study found that stroke patients who went on to have poor outcomes had substantially higher sCD163 levels (about 820 ng/mL on average) compared to those who recovered better (about 650 ng/mL). Importantly, this association held up even after accounting for other known risk factors like stroke severity scores, the size of the brain injury seen on imaging, and whether patients had an irregular heart rhythm.
The researchers developed a scoring model that combined sCD163 with these traditional clinical measures, and found that adding sCD163 meaningfully improved the model's ability to predict outcomes — increasing accuracy by about 12 percentage points. The model performed consistently well both in the original group of 664 patients and in an independent group of 300 patients from a different hospital, suggesting the findings are not specific to one institution or patient group. About one-third of patients in both groups experienced poor outcomes, defined as significant disability or worse on a standard stroke recovery scale.
This research suggests that routinely measuring sCD163 in stroke patients who do not receive clot-busting or clot-removal treatments could help clinicians identify high-risk individuals earlier and potentially tailor monitoring or rehabilitation efforts. However, the authors caution that further studies are needed to understand exactly how sCD163 contributes to poor recovery and to confirm these findings across more diverse patient populations.
Yan M, Shao L, Li Z, Liu C. (2026). Serum sCD163 as a potential biomarker for predicting poor functional outcome at 3 months in acute ischemic stroke without reperfusion: a prospective study.. Frontiers in immunology. https://doi.org/10.3389/fimmu.2026.1914293