Background Stroke is currently the second most common cause of death and disability-adjusted life years worldwide. However, there have been few comprehensive models reported for evaluating the short-term mortality of acute moderate and severe stroke patients. In this study, we aimed to investigate blood-based biomarkers of inflammation, immunity, nutritional metabolism and blood brain barrier damage at the early stages, explore the association of these biomarkers with 30-day mortality in acute moderate and severe stroke patients, and develop a nomogram specifically for 30-day mortality prediction in these patients.
Methods A prospective observational study was conducted, enrolling 152 acute stroke patients (including acute ischemic stroke (AIS) or intracerebral hemorrhage (ICH)) who had NIHSS >14 or GCS<8. The study was conducted between January 2020 and November 2022 in the neurological intensive care unit (NICU), and patients were consecutively enrolled. The laboratory parameters and clinical characteristics of the patients were collected by the research team. Blood biomarkers, including IL-10, MIP-1β, TNF-α, nNOS, iNOS, MMP-9, S-100β, and ET-1, were measured using ELISA within the first 24 hours following symptom onset. The least absolute shrinkage and selection operator (LASSO) regression optimized predictive clinical, biomarker, and combined models.Univariate and multivariate logistic regression analyses were performed to construct a nomogram model for predicting the 30-day mortality risk of acute moderate and severe stroke patients.The bootstrapping validation method, a resampling technique, was used to internally validate the nomogram model. The performance of the nomogram was evaluated on the basis of its calibration, discrimination, and clinical usefulness.
Results The 30-day mortality rate of acute moderate and severe stroke patients in the neurological intensive care unit (NICU) was 23.68%. Surviving stroke patients showed lower neutrophils, neutrophil-to-lymphocyte ratio (NLR), platelet count/platelet volume ratio (PPR), procalcitonin (PCT), IL-6, IL-10, TNF-α, and higher lymphocytes/monocytes ratio(LMR), lymphocytes than non-surviving patients, while there were no significant differences in monocyte, albumin, prealbumin, transferrin, MIP-1β, nNOS, iNOS, MMP-9, S-100β and ET-1 between non-surviving and surviving patients. The LASSO regression identified 3 variables (IL-10 (P<0.001), NIHSS (P=0.015) and cerebral herniation (P<0.001)), and a predictive model of 30-day mortality in acute moderate and severe stroke patients was subsequently established. The area under the curve (AUC) of the predictive model was 0.885 (95% CI: 0.808-0.963). The model achieved a concordance index of 0.877 [95% CI (0.775, 0.979)] and had a well-fitted calibration curve and good clinical application value.
Conclusions This study found that blood biomarkers, including neutrophils, NLR, PPR, PCT, IL-6, IL-10, TNF-α, LMR, and lymphocytes measured within the first 24 hours following symptom onset, were associated with 30-day mortality in acute moderate and severe stroke patients. Specifically, higher levels of IL-10 exhibited a greater predictive power.Furthermore, we constructed a predictive nomogram model, which may facilitate early prognosis identification, enhance communication between patients and clinicians, and improve patient management.