Abstract
Objective
Despite the rigorous application of guideline-directed medical therapy (GDMT) for secondary prevention, the residual risk of recurrent ischemic stroke (IS) remains substantial. This study aimed to evaluate the independent prognostic value of two novel composite biomarkers—the Monocyte-to-High-Density Lipoprotein Cholesterol Ratio (MHR) and the Neutrophil Percentage-to-Albumin Ratio (NPAR), and to determine whether integrating these biomarkers into a standard clinical risk model provides incremental predictive value over the Essen Stroke Risk Score (ESRS).
Methods
We conducted a retrospective cohort study of 907 patients with acute ischemic stroke admitted to Shanxi Bethune Hospital between January 2021 and December 2023. Patients were stratified into recurrence (n = 168) and non-recurrence (n = 739) groups based on follow-up (median 24.44 months; IQR: 20.57–31.29). We constructed hierarchical multivariable Cox proportional hazards models: Model A (clinical risk factors + discharge medications), Model B (Model A + MHR), and Model C (Model A + MHR + NPAR). The Essen Stroke Risk Score (ESRS) was calculated as a comparative baseline. Prognostic performance was assessed using Harrell’s C-index, Net Reclassification Improvement (NRI), and Integrated Discrimination Improvement (IDI). Time-dependent ROC analysis was utilized to assess the stability of the model’s predictive performance at 12 months, 24 months, and the end of follow-up. Multicollinearity was assessed using Variance Inflation Factors (VIF).
Results
The cumulative incidence of recurrence was 18.52%. In the fully adjusted Model C, both elevated MHR (HR = 1.43; 95% CI 1.22–1.67; P < 0.001) and NPAR (HR = 4.21; 95% CI 2.87–6.17; P < 0.001) were robust independent predictors of recurrence. Restricted cubic spline analysis revealed a J-shaped non-linear association for MHR (threshold > 0.628). The standard ESRS showed modest predictive ability (C-index = 0.608). In contrast, integrating MHR and NPAR into the clinical model significantly enhanced discrimination, raising the C-index from 0.801 (Model A) to 0.848 (Model C; P < 0.001). The combined model demonstrated substantial improvement in risk stratification (NRI = 0.547, P < 0.001; IDI = 0.088, P < 0.001) with low multicollinearity (VIF < 5 for all covariates). Time-dependent ROC analysis confirmed stable predictive accuracy with AUCs of 0.880 at 12 months and 0.883 at 24 months.
Conclusions
MHR and NPAR are robust, independent predictors of ischemic stroke recurrence that capture non-redundant biological risks not accounted for by the Essen Stroke Risk Score. Their integration into standard clinical models significantly improves prognostic accuracy, offering a readily accessible tool to identify high-risk patients who may benefit from intensified secondary prevention targeting residual inflammation.
Keywords: Ischemic stroke, Recurrence, Monocyte-to-High-Density lipoprotein cholesterol ratio, Neutrophil Percentage-to-Albumin ratio, Biomarkers, Risk stratification
Introduction
Ischemic stroke (IS) remains a leading cause of mortality and long-term disability worldwide, [1] with recurrence rates reaching 25–30% within five years despite advances in acute care [2, 3]. While secondary prevention strategies focus on controlling traditional risk factors through antithrombotics and statins, [4] a substantial “residual vascular risk” persists. Emerging evidence implicates the “lipid-inflammation axis” as a critical driver of this residual risk, [5] where chronic low-grade inflammation interacts with metabolic dysregulation to destabilize atherosclerotic plaques [6].
The Monocyte-to-High-Density Lipoprotein Cholesterol Ratio (MHR) has emerged as a novel composite biomarker reflecting this interplay [7, 8]. It quantifies the balance between pro-inflammatory monocytes, which drive plaque progression, [9] and high-density lipoprotein cholesterol (HDL-C), which exerts anti-inflammatory and reverse cholesterol transport effects [10]. Concurrently, the Neutrophil Percentage-to-Albumin Ratio (NPAR) serves as a marker of “systemic vulnerability,” capturing acute inflammatory intensity and compromised nutritional reserve [11].
Although elevated MHR and NPAR have been individually associated with poor stroke outcomes, [12, 13] no study has yet evaluated the incremental prognostic value of combining MHR and NPAR against a rigorously adjusted clinical baseline model that includes secondary prevention medications. Consequently, it remains unclear whether these biomarkers offer genuine incremental predictive value beyond established risk stratification tools like the Essen Stroke Risk Score (ESRS).
To bridge this knowledge gap, we investigated the independent association between baseline monocyte-to-high-density lipoprotein cholesterol ratio (MHR) and ischemic stroke recurrence, with additional evaluation of NPAR. We further characterized non-linear dose–response relationships and quantified the incremental prognostic value of incorporating these biomarkers into standard clinical models using net reclassification and integrated discrimination improvements to better identify residual inflammatory risk [14, 15].
Materials and Methods
Study Design and Population
We conducted a retrospective cohort study of patients admitted with acute ischemic stroke (AIS) to Shanxi Bethune Hospital between January 2021 and December 2023. The study protocol was formally approved by the Medical Ethics Committee of Shanxi Medical University (Approval No. 2025029). The conduct of this research adhered to the ethical principles outlined in the 1964 Declaration of Helsinki and its later amendments. Written informed consent was obtained from all participants or their legally authorized representatives during follow-up contact, prior to the inclusion of their clinical data in our research database. Participants for whom written consent could not be obtained were not included in the final analysis.
Inclusion and Exclusion Criteria
Eligible patients were aged ≥ 18 years with AIS confirmed by neuroimaging (CT or MRI) within 72 h of symptom onset. To rigorously isolate vascular inflammatory risk from acute infectious or systemic inflammatory confounding, we applied strict exclusion criteria: (1) active infection (e.g., pneumonia, urinary tract infection) or systemic inflammatory diseases within 2 weeks prior to admission; (2) severe hepatic or renal insufficiency; (3) hematological malignancies or autoimmune disorders; and (4) use of antibiotics or corticosteroids within 14 days of admission. Crucially, to reflect real-world secondary prevention, use of discharge secondary prevention medications (statins, antiplatelets, oral anticoagulants) was recorded and adjusted for in the analysis. Out of 1,080 screened patients, 907 met the final eligibility criteria. (Fig. 1).
Fig. 1.
STROBE Flowchart of Study Population Selection
Data Collection and Biomarker Definitions
Baseline demographic, clinical, and laboratory data were extracted from electronic medical records. Stroke severity was assessed using the National Institutes of Health Stroke Scale (NIHSS) [16] at admission. Stroke etiology was classified according to TOAST criteria. Secondary prevention medications were defined as discharge prescriptions at the end of the index hospitalization, including statins, antiplatelet agents, and oral anticoagulants. The Essen Stroke Risk Score (ESRS) [17] was calculated according to the original definition (0–9 points): age (< 65 years = 0, 65–75 years = 1, > 75 years = 2) and 1 point each for hypertension, diabetes mellitus, current smoking, previous myocardial infarction, other cardiovascular disease, peripheral arterial disease, prior stroke/TIA, and male sex.
Laboratory parameters were measured from fasting venous blood samples obtained within 24 h of admission. The MHR was calculated as the absolute monocyte count (×10⁹/L) divided by the HDL-C concentration (mmol/L). The NPAR was calculated as the neutrophil percentage (%) divided by the serum albumin concentration (g/L).
Outcome Assessment
The primary clinical endpoint was recurrent ischemic stroke, strictly defined as the sudden onset of new focal neurological deficits lasting > 24 h, with confirmation of a new acute ischemic lesion in a vascular territory distinct from the index stroke (or clearly new within the same territory) by CT or MRI. This definition excluded worsening of the initial deficit due to edema or hemorrhagic transformation. Follow-up was conducted via outpatient visits or telephone interviews performed by trained investigators until the censoring date of October 31, 2025, serving strictly as a method to ascertain clinical outcomes (i.e., stroke recurrence) without relying on any specific psychometric scale or newly developed questionnaire requiring validation.
Statistical Analysis
Analyses were performed using R software (version 4.5.2). Continuous variables were compared using the Student’s t-test or Mann-Whitney U test, and categorical variables using the Chi-square test. Cumulative incidence of recurrence was estimated using the Kaplan–Meier method. We constructed three hierarchical multivariable Cox proportional hazards models to assess incremental value: Model A (Clinical Baseline: age, sex, BMI, hypertension, diabetes, AF, prior stroke, smoking, NIHSS, TOAST, LDL-C, and discharge medications); Model B (Model A + MHR); and Model C (Model A + MHR + NPAR). The Essen Stroke Risk Score (ESRS) was utilized as an external reference model. Model discrimination was evaluated using Harrell’s C-index. The incremental predictive value was quantified using Net Reclassification Improvement (NRI) and Integrated Discrimination Improvement (IDI). We performed time-dependent ROC analysis to evaluate the Area Under the Curve (AUC) at 12 months, 24 months, and at the end of follow-up to assess the temporal stability of the model’s predictive power. Multicollinearity was assessed using Variance Inflation Factors (VIF), with a threshold of < 5 indicating no significant redundancy. Restricted cubic splines (RCS) with three knots were used to explore non-linear associations. A two-sided P-value < 0.05 was considered statistically significant.
Results
Baseline Characteristics of the Study Population
The final analysis included 907 patients with acute ischemic stroke. During the follow-up period, the median follow-up duration was 24.44 months (interquartile range, 20.57–31.29 months). A total of 168 patients (18.52%) experienced recurrent ischemic stroke. Detailed comparisons of demographic and clinical characteristics between the recurrence and non-recurrence groups are presented in Table 1. Patients in the recurrence group had significantly higher BMI, admission NIHSS scores, and prevalence of prior stroke (P < 0.05).
Table 2.
Baseline Laboratory Findings and Inflammatory Biomarkers
| Variables | Total Cohort (N = 907) |
Non-Recurrence (n = 739) |
Recurrence (n = 168) |
P Value |
|---|---|---|---|---|
| Blood Counts | ||||
| White blood cell count, ×10⁹/L | 7.56 (2.66) | 7.53 (2.66) | 7.65 (2.69) | 0.614 |
| Neutrophil count, ×10⁹/L | 5.26 (2.39) | 5.21 (2.41) | 5.45 (2.28) | 0.247 |
| Lymphocyte count, ×10⁹/L | 1.50 (1.18) | 1.57 (1.27) | 1.17 (0.57) | < 0.001 |
| Monocyte count, ×10⁹/L | 0.63 (0.67) | 0.65 (0.72) | 0.61 (0.33) | 0.209 |
| Biochemical Parameters | ||||
| Fasting blood glucose, mmol/L | 6.00 (5.40–8.10) | 6.00 (5.40-8.00) | 6.50 (5.30–9.03) | 0.101 |
| Albumin, g/L | 36.81 (4.66) | 37.06 (4.53) | 35.70 (5.04) | 0.001 |
| Total cholesterol (TC), mmol/L | 3.54 (0.99) | 3.53 (1.01) | 3.60 (0.93) | 0.397 |
| Triglycerides (TG), mmol/L | 1.42 (1.12–1.77) | 1.41 (1.12–1.77) | 1.44 (1.19–1.74) | 0.754 |
| LDL-C, mmol/L | 2.27 (0.80) | 2.21 (0.80) | 2.54 (0.79) | < 0.001 |
| HDL-C, mmol/L | 0.94 (0.26) | 0.96 (0.26) | 0.90 (0.25) | 0.160 |
| CRP, mg/L, Median (IQR) | 7.55 (2.04–26.27) | 5.33 (1.96–22.15) | 20.55 (5.49–39.39) | < 0.001 |
| Novel Inflammatory Indices | ||||
| MHR (Monocyte/HDL-C ratio) | 0.63 (0.44–0.88) | 0.59 (0.43–0.81) | 0.87 (0.62–1.33) | < 0.001 |
| NPAR (Neutrophil%/Albumin) | 1.82 (1.60–2.10) | 1.77 (1.55–2.04) | 2.08 (1.80–2.34) | < 0.001 |
CRP, MHR, NPAR, Glucose, TG are presented as median (IQR) due to skewed distribution. Other continuous variables are mean (SD)
Table 1.
Baseline Demographic and Clinical Characteristics
| Variables | Total Cohort (N = 907) |
Non-Recurrence (n = 739) |
Recurrence (n = 168) |
P Value |
|---|---|---|---|---|
| Demographics | ||||
| Age, years, mean (SD) | 62.29 (12.16) | 62.14 (12.37) | 62.96 (11.21) | 0.429 |
| Male sex, n (%) | 641 (70.7) | 518 (70.1) | 123 (73.2) | 0.479 |
| BMI, kg/m², mean (SD) | 25.20 (3.40) | 25.07 (3.22) | 25.77 (4.07) | 0.015 |
| Smoking history, n (%) | 484 (53.4) | 395 (53.5) | 89 (53.0) | 0.980 |
| Alcohol consumption, n (%) | 378 (41.7) | 297 (40.2) | 81 (48.2) | 0.069 |
| Comorbidities | ||||
| Hypertension, n (%) | 616 (67.9) | 501 (67.8) | 115 (68.5) | 0.942 |
| Diabetes mellitus, n (%) | 266 (29.3) | 208 (28.1) | 58 (34.5) | 0.122 |
| Atrial fibrillation, n (%) | 146 (16.1) | 121 (16.4) | 25 (14.9) | 0.720 |
| Prior Stroke, n (%) | 195 (21.5) | 131 (17.7) | 64 (38.1) | < 0.001 |
| Clinical Assessment | ||||
| Systolic BP, mmHg, mean (SD) | 130.84 (16.16) | 130.71 (16.21) | 131.45 (15.94) | 0.593 |
| Admission NIHSS score, median (IQR) | 5.00 (3.00–6.00) | 5.00 (3.00–5.00) | 6.00 (5.00–7.00) | < 0.001 |
| ESRS, median (IQR) | 2.00 (2.00–3.00) | 2.00 (1.00–3.00) | 3.00 (2.00–4.00) | < 0.001 |
| TOAST Classification, n (%) | 0.014 | |||
| Large-artery atherosclerosis (LAA) | 410 (45.2) | 319 (43.2) | 91 (54.2) | |
| Cardioembolism (CE) | 154 (17.0) | 121 (16.4) | 33 (19.6) | |
| Small-vessel occlusion (SVO) | 230 (25.4) | 203 (27.5) | 27 (16.1) | |
| Other determined etiology (SOE) | 71 (7.8) | 60 (8.1) | 11 (6.5) | |
| Undetermined etiology (SUE) | 42 (4.6) | 36 (4.9) | 6 (3.6) | |
| Length of stay, days, median (IQR) | 15.0 (13.0–17.0) | 15.0 (13.0–17.0) | 16.0 (14.0–19.0) | < 0.001 |
| Secondary Prevention Medications | ||||
| Statin Use, n (%) | 809 (89.2) | 690 (93.4) | 119 (70.8) | < 0.001 |
| Antiplatelet Use, n (%) | 704 (77.6) | 589 (79.7) | 115 (68.5) | 0.002 |
| Anticoagulant Use, n (%) | 116 (12.8) | 105 (14.2) | 11 (6.5) | 0.011 |
The Essen Stroke Risk Score (ESRS), calculated for the entire cohort, showed a median score of 3 (IQR 2–4) in the recurrence group compared to 2 (IQR 1–3) in the non-recurrence group (P < 0.001). Notably, the recurrence group had significantly lower rates of statin, antiplatelet, and anticoagulant use (P < 0.05), highlighting the protective role of secondary prevention medications. In terms of stroke etiology, Large-Artery Atherosclerosis (LAA) was more prevalent in the recurrence group.
Table 2 summarizes laboratory findings. The recurrence group exhibited distinct inflammatory and metabolic profiles, characterized by significantly higher LDL-C, CRP, and lower serum albumin (P < 0.001). Crucially, both MHR (0.87 vs. 0.59) and NPAR (2.08 vs. 1.77) were significantly elevated in patients who suffered recurrence.
Survival Analysis and Dose-Response Relationship
Kaplan–Meier analysis stratified by MHR quartiles (Fig. 2) demonstrated a significant graded association, with patients in the highest MHR quartile (Q4) exhibiting the steepest increase in cumulative recurrence incidence compared with the lower quartiles (log-rank test P < 0.001).
Fig. 2.
Cumulative incidence estimates of ischemic stroke recurrence stratified by MHR quartiles
To further characterize the relationship between MHR and recurrence risk, we conducted restricted cubic spline (RCS) analysis. As shown in Fig. 3, a non-linear, J-shaped dose-response relationship was observed (P for non-linearity < 0.001). The risk of recurrence remained relatively stable and low at MHR levels below 0.628. However, beyond this threshold, the hazard ratio increased exponentially, eventually plateauing at MHR values > 1.247. This defines a specific “high-risk biological phenotype” characterized by MHR > 0.628.
Fig. 3.
Dose–response association between MHR and risk of ischemic stroke recurrence based on restricted cubic spline analysis
Predictors of Ischemic Stroke Recurrence
Univariate Cox regression (Table 3) identified several significant predictors of recurrence, including BMI, prior stroke, admission NIHSS, LAA subtype, LDL-C, MHR, and NPAR. Conversely, the use of statins, antiplatelets, and anticoagulants was associated with a significant protective effect, confirming the efficacy of standard care.
Table 3.
Univariate Cox Proportional Hazards Regression Analysis for Stroke Recurrence
| Variables | HR (95% CI) | P Value |
|---|---|---|
| Age | 1.01 (0.99–1.02) | 0.338 |
| Male sex | 1.14 (0.81–1.61) | 0.442 |
| BMI | 1.05 (1.01–1.10) | 0.019 |
| Smoking history | 1.01 (0.75–1.37) | 0.942 |
| Alcohol consumption | 1.33 (0.98–1.80) | 0.065 |
| Hypertension | 1.02 (0.74–1.42) | 0.895 |
| Diabetes mellitus | 1.31 (0.95–1.80) | 0.096 |
| Atrial fibrillation | 0.92 (0.60–1.41) | 0.712 |
| Prior Stroke | 2.63 (1.93–3.60) | < 0.001 |
| Admission NIHSS score | 1.47 (1.35–1.59) | < 0.001 |
| ESRS | 1.48 (1.25–1.76) | < 0.001 |
| TOAST (LAA type) | 1.51 (1.11–2.04) | 0.008 |
| Statin Use | 0.24 (0.17–0.33) | < 0.001 |
| Antiplatelet Use | 0.58 (0.42–0.81) | 0.001 |
| Anticoagulant Use | 0.46 (0.25–0.85) | 0.013 |
| LDL-C, mmol/L | 1.55 (1.29–1.85) | < 0.001 |
| MHR | 1.55 (1.37–1.75) | < 0.001 |
| NPAR | 6.08 (4.19–8.82) | < 0.001 |
To evaluate the independent and incremental value of the biomarkers, we constructed three hierarchical multivariable Cox proportional hazards models (Table 4). Before finalizing these models, we performed a multicollinearity diagnostic. MHR (VIF = 1.08) and NPAR (VIF = 1.11) showed no significant redundancy with LDL-C or other clinical variables (all VIF < 5), validating their inclusion as independent covariates. In the fully adjusted Model C, both MHR (HR = 1.43; 95% CI 1.22–1.67; P < 0.001) and NPAR (HR = 4.21; 95% CI 2.87–6.17; P < 0.001) remained robust independent predictors of recurrence, even after adjusting for secondary prevention medications and stroke subtypes.
Table 4.
Multivariate Cox Proportional Hazards Regression Analysis for Stroke Recurrence
| Variables | Model A | Model B | Model C | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | P Value | HR (95% CI) | P Value | HR (95% CI) | P Value | |
| Primary Predictors | ||||||
| MHR | 1.42 (1.23–1.65) | < 0.001 | 1.43 (1.22–1.67) | < 0.001 | ||
| NPAR | 4.21 (2.87–6.17) | < 0.001 | ||||
| Adjusted Covariates | ||||||
| Age | 1.01 (0.99–1.02) | 0.299 | 1.01 (0.99–1.02) | 0.226 | 0.99 (0.98–1.01) | 0.439 |
| Male Sex | 1.50 (0.99–2.28) | 0.054 | 1.48 (0.98–2.25) | 0.063 | 1.22 (0.81–1.83) | 0.349 |
| BMI | 1.06 (1.02–1.11) | 0.006 | 1.06 (1.02–1.11) | 0.004 | 1.05 (1.00–1.09) | 0.047 |
| Hypertension | 0.91 (0.64–1.28) | 0.577 | 0.90 (0.64–1.28) | 0.562 | 0.98 (0.68–1.39) | 0.897 |
| Diabetes mellitus | 0.95 (0.67–1.35) | 0.787 | 0.98 (0.69–1.39) | 0.907 | 1.13 (0.80–1.61) | 0.489 |
| Prior Stroke | 2.34 (1.68–3.24) | < 0.001 | 2.24 (1.61–3.12) | < 0.001 | 1.91 (1.36–2.67) | < 0.001 |
| Atrial Fibrillation | 1.59 (0.85–2.97) | 0.146 | 1.48 (0.79–2.77) | 0.216 | 1.58 (0.86–2.93) | 0.143 |
| Smoking history | 0.85 (0.59–1.22) | 0.377 | 0.87 (0.60–1.26) | 0.475 | 0.99 (0.69–1.42) | 0.956 |
| LDL-C | 1.61 (1.31–1.96) | < 0.001 | 1.63 (1.33–1.99) | < 0.001 | 1.49 (1.24–1.81) | < 0.001 |
| NIHSS Score | 1.42 (1.31–1.55) | < 0.001 | 1.39 (1.28–1.52) | < 0.001 | 1.38 (1.27–1.50) | < 0.001 |
| TOAST (LAA) | 1.47 (1.06–2.04) | 0.023 | 1.45 (1.04–2.01) | 0.028 | 1.52 (1.09–2.12) | 0.014 |
| Statin Use | 0.30 (0.21–0.43) | < 0.001 | 0.30 (0.21–0.43) | < 0.001 | 0.33 (0.23–0.47) | < 0.001 |
| Antiplatelet Use | 0.42 (0.28–0.62) | < 0.001 | 0.41 (0.28–0.61) | < 0.001 | 0.44 (0.30–0.65) | < 0.001 |
| Anticoagulant Use | 0.26 (0.12–0.61) | 0.002 | 0.29 (0.13–0.67) | 0.004 | 0.27 (0.12–0.64) | 0.003 |
Model A was adjusted for age, sex, BMI, smoking, hypertension, diabetes, prior stroke, atrial fibrillation, LDL-C, NIHSS, TOAST classification, and discharge medications (statins, antiplatelet agents, and anticoagulants). Model B included Model A plus MHR. Model C included Model A plus MHR and NPAR
Incremental Predictive Value and Model Comparison
As shown in Table 5, the inclusion of MHR in the clinical baseline model (Model B) resulted in a further improvement in predictive performance, with a C-index of 0.815 (95% CI, 0.795–0.835) and a ΔC-index of 0.207 (95% CI, 0.164–0.252). The proportional hazards assumption was not violated (PH test, P = 0.512). Model B also demonstrated enhanced risk reclassification and discrimination compared with ESRS, as reflected by a continuous NRI of 0.459 (95% CI, 0.294–0.627) and an IDI of 0.073 (95% CI, 0.049–0.098).
Table 5.
Incremental Predictive Performance of Clinical and Biomarker-Enhanced Models Compared with ESRS
| Model Structure | C-index (95% CI) | ΔC-index (95% CI) | PH Test | NRI (95% CI) | IDI (95% CI) |
|---|---|---|---|---|---|
| ESRS | 0.608 (0.567–0.655) | ||||
| Model A | 0.801 (0.781–0.821) | 0.193 (0.149–0.240) | 0.428 | 0.378 (0.208–0.544) | 0.061 (0.038–0.085) |
| Model B | 0.815 (0.795–0.835) | 0.207 (0.164–0.252) | 0.512 | 0.459 (0.294–0.627) | 0.073 (0.049–0.098) |
| Model C | 0.848 (0.829–0.867) | 0.240 (0.193–0.285) | 0.634 | 0.547 (0.375–0.712) | 0.088 (0.061–0.116) |
ΔC-index values represent improvement relative to the ESRS model
Further addition of NPAR (Model C) yielded the highest predictive performance, achieving a C-index of 0.848 (95% CI, 0.829–0.867) and a ΔC-index of 0.240 (95% CI, 0.193–0.285), with preserved proportional hazards (P = 0.634). Model C showed the greatest improvement in risk reclassification and discrimination (NRI, 0.547; 95% CI, 0.375–0.712; IDI, 0.088; 95% CI, 0.061–0.116).
Time-dependent ROC analysis was conducted to assess the stability of the combined model’s predictive performance across different follow-up horizons. Model C maintained robust discriminative power throughout the study period, with an AUC of 0.880 at 12 months and 0.883 at 24 months. Furthermore, at the termination of follow-up, the model sustained a high predictive accuracy (Global AUC = 0.878). This sustained accuracy across the 12-month, 24-month, and study-end timepoints suggests that baseline elevations in MHR and NPAR reflect a chronic, persistent inflammatory vulnerability rather than a transient acute-phase response (Fig. 4).
Fig. 4.
Comparison of time-dependent ROC curves for predicting ischemic stroke recurrence
Discussion
Our analysis indicates that baseline elevations in MHR and NPAR are independently associated with ischemic stroke recurrence and provide significant incremental prognostic value beyond conventional clinical risk factors and the Essen Stroke Risk Score (ESRS). Incorporation of these biomarkers into a comprehensive clinical model substantially improved risk stratification (C-index 0.848, NRI 0.547), enabling identification of patients with high residual inflammatory risk who may be overlooked by traditional scoring systems.
The superior discriminative performance of the combined model highlights the limitations of relying solely on static clinical history for prognostication. Although the ESRS incorporates established vascular risk factors, it may not fully capture the dynamic biological activity underlying atherosclerotic disease [18]. Our findings are consistent with the hypothesis that residual risk reflects the interplay between lipid metabolism and innate immunity, commonly referred to as the “lipid-inflammation axis” [19].
The independent prognostic value of MHR observed in our cohort lends clinical support to the “tipping point” theory of atherogenesis [20]. Biologically, this ratio functions as a barometer of plaque instability [21]. Monocytes serve as the obligate precursors for lesional macrophages; their recruitment and subsequent uptake of oxidized LDL via scavenger receptors (e.g., CD36) are fundamental steps in foam cell formation and necrotic core expansion [22]. Under homeostatic conditions, HDL-C counters this process by facilitating cholesterol efflux via ABCA1 transporters and suppressing the endothelial expression of adhesion molecules [23]. An elevated MHR, therefore, reflects a distinct pathophysiological state where pro-inflammatory monocyte activity overwhelms the protective and anti-oxidative capacity of HDL-C [20]. This imbalance implies the presence of biologically active, inflamed plaques that are prone to rupture, potentially explaining the higher recurrence risk observed in patients with elevated MHR despite comparable LDL-C levels [24].
Complementing this, the predictive utility of NPAR highlights the critical role of systemic immunothrombotic vulnerability [25]. Neutrophils are not merely markers of inflammation but active participants in the thrombotic process [26]. Upon activation, neutrophils release Neutrophil Extracellular Traps (NETs)—web-like chromatin structures that bind platelets and erythrocytes, promote fibrin deposition, and exhibit resistance to fibrinolysis [27]. This phenomenon of “immunothrombosis” creates a pro-coagulant environment that may precipitate recurrent occlusion [28]. Conversely, serum albumin functions as a potent antioxidant and neuroprotectant, scavenging reactive oxygen species via its cysteine-34 residue and maintaining microvascular integrity [29]. The strong association between elevated NPAR and recurrence (HR 4.21) suggests that the convergence of heightened neutrophil-driven thrombosis and depleted physiological reserve creates a fertile soil for secondary ischemic events [30].
Collectively, these findings support the concept of “Residual Inflammatory Risk” (RIR) as a distinct therapeutic target. While statins and antiplatelets address lipid levels and platelet aggregation, they may not fully resolve the upstream inflammatory signaling driven by the monocyte-neutrophil axis [31]. The significant reclassification improvement observed in our study implies that a dual-biomarker approach could serve as an accessible, low-cost screening tool to identify high-risk patients who might benefit from novel anti-inflammatory therapies (e.g., colchicine) or intensified lipid-lowering strategies [32].
Limitations
Several limitations must be acknowledged. First, as a single-center retrospective study, the potential for selection bias exists, although we rigorously included patients on secondary prevention medications to reflect real-world practice. Second, our study database did not capture specific causes of death during follow-up. Consequently, we were unable to perform a competing risk analysis (e.g., Fine-Gray model) to distinguish non-stroke-related mortality from censoring. Third, single baseline measurements do not capture the dynamic evolution of inflammatory status over time. Finally, while we compared performance against ESRS, external validation in diverse cohorts is necessary to generalize these thresholds.
Conclusion
Elevated MHR and NPAR are significantly associated with ischemic stroke recurrence, reflecting distinct pathways of residual inflammatory and systemic risk. Their integration into standard clinical models significantly outperforms the Essen Stroke Risk Score, enhancing prognostic accuracy and risk reclassification. This dual-biomarker approach provides a practical, low-cost strategy to identify high-risk patients who require intensified secondary prevention beyond standard guideline-directed therapy.
Acknowledgements
The authors would like to express their sincere gratitude to all participants for their cooperation and willingness to take part in this study. We are also grateful to the staff of the Department of Rehabilitation Medicine, Shanxi Bethune Hospital, for their valuable assistance with data collection.
Authors’ contributions
ZY conceived and designed the study, conducted data collection and statistical analyses, interpreted the results, and drafted the main manuscript text and figures. PQ contributed to patient recruitment, clinical data acquisition, interpretation of clinical findings, and critical revision of the manuscript. MZ assisted with data management, statistical support, construction of tables, and refinement of the Methods and Results sections. YL contributed to the literature review, data cleaning, and drafting and revising parts of the Introduction and Discussion. PW and LZ critically reviewed and revised the manuscript for important intellectual content, and secured project funding. All authors read and approved the final manuscript and agree to be accountable for all aspects of the work.
Funding
This work was supported by the Key Research and Development Program Project of Shanxi Province (Grant No. 202302130501014) and the Shanxi Provincial Department of Science and Technology (Grant No. 202302130501004). The authors declare that no external commercial funding or industry-specific grants were received for this study.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available due to patient confidentiality requirements but are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The study protocol was formally approved by the Medical Ethics Committee of Shanxi Medical University (Approval No. 2025029). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants or their legally authorized representatives.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Liangyuan Zhao, Email: 414861955@qq.com.
Pingzhi Wang, Email: wpzcxl@163.com.
References
- 1.Li X, Kong X, Yang C, et al. Global, regional, and National burden of ischemic stroke, 1990–2021: an analysis of data from the global burden of disease study 2021[J]. EClinicalMedicine. 2024;75:102758. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tu WJ, Chao BH, Ma L, et al. Case-fatality, disability and recurrence rates after first-ever stroke: a study from bigdata observatory platform for stroke of China[J]. Brain Res Bull. 2021;175:130–5. [DOI] [PubMed] [Google Scholar]
- 3.de Havenon A, Viscoli C, Kleindorfer D, et al. Disability and recurrent stroke among participants in stroke prevention Trials[J]. JAMA Netw Open. 2024;7(7):e2423677–2423677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kleindorfer DO, Towfighi A, Chaturvedi S, et al. 2021 guideline for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline from the American heart Association/American stroke Association[J]. Stroke. 2021;52(7):e364–467. [DOI] [PubMed] [Google Scholar]
- 5.Ridker PM. Targeting residual inflammatory risk: the next frontier for atherosclerosis treatment and prevention[J]. Vascul Pharmacol. 2023;153:107238. [DOI] [PubMed] [Google Scholar]
- 6.Libby P. The changing landscape of atherosclerosis[J]. Nature. 2021;592(7855):524–33. [DOI] [PubMed] [Google Scholar]
- 7.Nam KW, Kwon HM, Jeong HY, et al. Monocyte to high-density lipoprotein cholesterol ratio is associated with cerebral small vessel diseases[J]. BMC Neurol. 2024;24(1):18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Arabi A, Abdelhamid A, Nasrallah D, et al. Monocyte-to-HDL ratio (MHR) as a novel biomarker: reference ranges and associations with inflammatory diseases and disease-specific mortality[J]. Lipids Health Dis. 2025;24(1):343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kong P, Cui ZY, Huang XF, et al. Inflammation and atherosclerosis: signaling pathways and therapeutic intervention[J]. Signal Transduct Target Therapy. 2022;7(1):131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rohatgi A, Westerterp M, Von Eckardstein A, et al. HDL in the 21st century: a multifunctional roadmap for future HDL research[J]. Circulation. 2021;143(23):2293–309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lu Y, Mao B, Wang M, et al. Neutrophil Percentage-to-Albumin ratio as a prognostic marker for mortality in ischemic stroke Patients[J]. Int J Med Sci. 2025;22(11):2663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li Y, Xi L, Li L. Predictive value of monocyte-to-high-density lipoprotein-cholesterol ratio (MHR) for poor prognosis after intravenous thrombolytic therapy for acute ischaemic stroke[J]. Archives Med Science: AMS. 2024;20(6):1809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Deng C, Liu B, Wang M, et al. Analysis of the correlation between neutrophil percentage-to-albumin ratio, neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio with short-term prognosis in acute ischemic stroke patients undergoing intravenous thrombolysis[J]. Front Neurol. 2025;16:1512355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Li J, Pan Y, Xu J, et al. Residual inflammatory risk predicts poor prognosis in acute ischemic stroke or transient ischemic attack Patients[J]. Stroke. 2021;52(9):2827–36. [DOI] [PubMed] [Google Scholar]
- 15.Desai SR, Dhindsa DS, Ko YA, et al. Aggregate clinical and biomarker-based model predicts adverse outcomes in patients with coronary artery disease[J]. Am J Cardiol. 2023;203:315–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Brott T, Adams HP Jr, Olinger CP, et al. Measurements of acute cerebral infarction: a clinical examination scale[J]. Stroke. 1989;20(7):864–70. [DOI] [PubMed] [Google Scholar]
- 17.Weimar C, Diener HC, Alberts MJ, et al. The Essen stroke risk score predicts recurrent cardiovascular events: a validation within the reduction of atherothrombosis for continued health (REACH) registry[J]. Stroke. 2009;40(2):350–4. [DOI] [PubMed] [Google Scholar]
- 18.Xiong Y, Wang S, Li Z, et al. Thirteen-year trends in risk scores predictive values for subsequent stroke in patients with acute ischemic event[J]. Brain Behav. 2023;13(5):e2962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Mensah GA, Arnold N, Prabhu SD, Ridker PM, Welty FK. Inflammation and cardiovascular disease: 2025 ACC scientific statement: A report of the American college of cardiology. J Am Coll Cardiol 2025 Sep 29:S0735-1097(25)07555-2. [DOI] [PubMed]
- 20.Li Y, Chen D, Sun L, et al. Monocyte/high-density lipoprotein ratio predicts the prognosis of large artery atherosclerosis ischemic stroke[J]. Front Neurol. 2021;12:769217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ajoolabady A, Pratico D, Lin L, et al. Inflammation in atherosclerosis: pathophysiology and mechanisms[J]. Volume 15. Cell death & disease; 2024. p. 817. 11. [DOI] [PMC free article] [PubMed]
- 22.Tabas I, Bornfeldt KE. Intracellular and intercellular aspects of macrophage immunometabolism in atherosclerosis[J]. Circul Res. 2020;126(9):1209–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ogura M. HDL, cholesterol efflux, and ABCA1: free from good and evil dualism[J]. J Pharmacol Sci. 2022;150(2):81–9. [DOI] [PubMed] [Google Scholar]
- 24.Gui Y, Zheng H, Cao RY. Foam cells in atherosclerosis: novel insights into its origins, consequences, and molecular mechanisms[J]. Front Cardiovasc Med. 2022;9:845942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Middleton EA, He XY, Denorme F, et al. Neutrophil extracellular traps contribute to immunothrombosis in COVID-19 acute respiratory distress syndrome[J]. Blood J Am Soc Hematol. 2020;136(10):1169–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Shetty S, Subramanian M. Neutrophil extracellular traps (NETs) as drivers of atherosclerosis: pathogenic mechanisms and therapeutic opportunities[J]. Pharmacology & Therapeutics; 2025. p. 108917. [DOI] [PubMed]
- 27.Xu X, Wu Y, Xu S, et al. Clinical significance of neutrophil extracellular traps biomarkers in thrombosis[J]. Thromb J. 2022;20(1):63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Zaiema S, E D G E D, Elwafa M A Z M A A, Hassan SGA, et al. Insight into antiphospholipid syndrome: the role and clinical utility of neutrophils extracellular traps formation[J]. Thromb J. 2024;22(1):32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Tabata F, Wada Y, Kawakami S, et al. Serum albumin redox states: more than oxidative stress biomarker[J]. Antioxidants. 2021;10(4):503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yang D, Niu C, Li P, et al. Study of the neutrophil percentage-to-albumin ratio as a biomarker for predicting recurrence of first-episode ischemic stroke[J]. J Stroke Cerebrovasc Dis. 2024;33(1):107485. [DOI] [PubMed] [Google Scholar]
- 31.Mohammadnia N, Opstal TSJ, El Messaoudi S, et al. An update on inflammation in atherosclerosis: how to effectively treat residual risk[J]. Clin Ther. 2023;45(11):1055–9. [DOI] [PubMed] [Google Scholar]
- 32.Pathangey G, Allam MN, Abdelnabi MH, et al. Beyond lipids and platelets: A review of Anti-Inflammatory strategies in secondary prevention of acute coronary Syndromes[J]. J Clin Med. 2025;14(22):7910. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available due to patient confidentiality requirements but are available from the corresponding author on reasonable request.