アテローム性動脈硬化性心血管疾患における炎症および代謝バイオマーカーを用いたリスク層別化：死亡率に関する多コホート予測研究

The pathological process of atherosclerotic cardiovascular disease (ASCVD) involves complex interactions between metabolic dysregulation and inflammatory responses. The stress hyperglycemia ratio (SHR) and neutrophil-to-lymphocyte ratio (NLR), as biomarkers reflecting metabolic stress and systemic inflammation respectively, have demonstrated significant value in ASCVD prognosis assessment. This study aims to investigate the predictive role of combined SHR and NLR indicators for all-cause mortality in ASCVD patients and their clinical applicability.

ASCVD patients were identified from the Medical Information Mart for Intensive Care IV (MIMIC-IV) and National Health and Nutrition Examination Survey (NHANES) databases, stratified by SHR/NLR tertiles. Multivariable Cox regression, restricted cubic splines, time-dependent ROC, and machine learning models assessed mortality associations.

Among 6159 patients, the highest SHR tertile showed increased mortality (NHANES: HR = 1.20, 95 % CI 1.08-1.34; MIMIC-IV: HR = 2.27, 95 % CI 1.44-3.57). The highest NLR tertile showed elevated risk (NHANES: HR = 1.56, 95 % CI 1.39-1.74; MIMIC-IV: HR = 1.61, 95 % CI 1.03-2.52). The combined high SHR/NLR group exhibited the most pronounced risk (NHANES: HR = 1.54, 95 % CI 1.35-1.77; MIMIC-IV: HR = 2.90, 95 % CI 1.70-4.93), with significantly improved predictive accuracy.

The combined SHR-NLR assessment effectively quantifies combined metabolic-inflammatory injury and provides superior prognostic stratification for ASCVD patients compared to individual biomarker evaluation. These findings highlight the clinical potential of this dual-biomarker approach for enhancing risk prediction, though further prospective validation is warranted to establish its predictive utility across various patient populations.