Neuroimmunoinflammatory axis and stroke severity as predictors of infection associated with acute ischemic stroke

Abstract

Introduction: stroke-associated infection is a frequent complication after ischemic stroke and is associated with worse functional prognosis and increased mortality. Acute brain injury can induce systemic immune-inflammatory dysregulation that increases susceptibility to infections. Objective: to evaluate the predictive value of stroke severity and its systemic immune-inflammatory response for the early detection of infection associated with acute ischemic stroke.

Methods: a prospective observational analytical cohort study was conducted, including 1176 patients with acute ischemic stroke. A logistic regression model was built to identify independent predictors. Model performance was assessed using the Hosmer-Lemeshow test, Nagelkerke R², and discriminative ability through the area under the ROC curve.

Results: the final model retained 5 independent predictors; an NIHSS score ≥ 12 was associated with the highest probability of infection (OR 10.0; 95 % CI: 7.15-14.24). Other independent predictors were absolute lymphocyte values < 2.2 x10⁹/L (OR 3.6; 95 % CI: 2.59-5.19), creatinine ≥ 105 μmol/L (OR 2.7; 95 % CI: 1.96-3.93), blood glucose ≥ 6.0 mmol/L (OR 1.8; 95 % CI: 1.32-2.64), and age ≥ 70 years (OR 1.8; 95 % CI: 1.29-2.56). The area under the ROC curve was 0.86, with a specificity of 94.33 %.

Conclusions: neurological severity, combined with markers of systemic immunoinflammatory dysfunction, allows predicting the risk of stroke-associated infection. This model could facilitate early risk stratification.