Data Management And Sharing System of China National Clinical Research Center for Neurological Diseases

Journal：J Am Heart Assoc

Publication type: Original Contributions

IF(IF): 5.3

PMID: 41614304 Link To Pubmed Full text Download

Stroke remains the third most common cause of death and the fourth most common cause of disability-adjusted life years globally, with poor functional outcomes posing a substantial threat to human health and economies. Inflammation and malnutrition have emerged as critical contributors to the development and progression of stroke. Circulating biomarkers that reflect inflammatory and nutritional status, such as serum albumin (ALB), total cholesterol (TC), and immune cell ratios, have been extensively studied for their potential predictive value in the prognosis of stroke. However, most of these markers were studied individually, lacking a comprehensive approach to assessing the combined effects of nutrition and inflammation in stroke patients. The Naples Prognostic Score (NPS) is a novel scoring system that integrates multiple indicators of nutritional, inflammatory, and immune status to predict clinical outcomes. The parameters of the NPS include ALB, TC, neutrophil-to-lymphocyte ratio (NLR), and lymphocyte-to-monocyte ratio (LMR), offering a more holistic assessment of both the inflammatory and nutritional status. Originally developed for cancer patients, the NPS has recently been explored to estimate the likelihood of developing diverse clinical conditions. Evidence suggests that the NPS is significantly associated with in-hospital mortality in ischemic stroke patients undergoing endovascular therapy, and may serve as a reliable marker for predicting stroke prevalence and long-term mortality in adults. However, associations of the NPS with functional outcomes in patients with acute ischemic stroke (AIS) or transient ischemic attack (TIA) remain unclear. Aging, characterized by the gradual deterioration of multiple biological systems, is another key factor influencing stroke prognosis. Biological age (BA), reflecting an individual's functional state rather than their chronological age, can be estimated by conventional clinical parameters. A previous study has shown that BA, especially PhenoAge acceleration (PAA), is associated with the prognosis of patients with AIS or TIA. Chronic, low-grade systemic inflammation and malnutrition are key mechanisms leading to BA. Given the intertwined roles of inflammation, nutrition, and aging, we hypothesized that PAA may mediate the associations between NPS and post-stroke outcomes.

Continuous variables were reported as medians and interquartile ranges (IQR) due to skewed distribution, and categorical variables were reported as frequencies with percentages. The differences in baseline characteristics between groups were compared using Kruskal-Wallis or Wilcoxon test for continuous variables and chi‐square test or Fisher's exact test for categorical variables. We took the low-NPS group as the reference and performed logistic regression to access the associations between the NPS scores and poor functional outcomes at 3 months and 1 year. Since patients were recruited from 201 hospitals, we treated hospitals as clusters and applied a robust sandwich variance estimator to adjust for intra-cluster correlation. Odds ratios (ORs) with 95% confidence intervals (CIs) were reported to indicate the strength of associations. We adjusted the baseline potential confounders and fitted two adjusted models. Model 1 adjusted for age and sex; Model 2 additionally adjusted for BMI, low-density lipoprotein cholesterol, current smoker, current alcohol drinking, medical history (stroke/TIA, diabetes, dyslipidemia, coronary heart disease, atrial fibrillation, and heart failure), TOAST, stroke type, prestroke mRS score, medication in hospital (antiplatelet, anticoagulant, cholesterol‐lowering, and hypoglycemic agents) and the NIHSS score at admission. Additionally, we combined the low and medium NPS groups and assessed the joint effects of NPS and PAA by categorizing patients into four groups based on NPS (low-to-medium vs high) and PAA (< 0 vs ≥ 0). In sensitivity analysis, we performed multiple imputation using chained equations to address missing data and reassessed the associations accordingly. The incremental predictive value of adding NPS to the basic model was evaluated using C-statistics, integrated discrimination improvement (IDI), and net reclassification index (NRI). Subgroup analyses were performed for subgroups of age, sex, BMI, baseline NIHSS score, dyslipidemia, diabetes mellitus, and coronary heart disease. To examine whether the associations between NPS and poor functional outcomes would be mediated by PAA, we performed a causal mediation analysis based on the method proposed by VanderWeele. The causal model included NPS as the predictor (X), PAA as the mediator (M), and poor functional outcomes at 3 months and 1 year as the outcomes (Y). The total effect was decomposed into the natural direct effect and the natural indirect effect (mediated via PAA), both expressed as ORs. The mediation effect (%) was measured by the proportion of the total effect mediated through the mediators, calculated as natural indirect effect/total effect × 100% on the log‐transformed OR scale. Confounders from Model 2 were adjusted in the mediation analysis.