Abstract: Objective To construct and validate a predictive model for the prognosis of acute ischemic stroke in the elderly based on inflammatory and nutritional indicators. Methods A total of 586 elderly patients with acute ischemic stroke from November 2021 to May 2022 in a tertiary grade-A hospital in Tangshan City were divided into good prognosis group (n=438) and poor prognosis group (n=148). Risk factors were analyzed by LASSO regression with binary logistic regression, and a nomogram was developed and validated. Results Multifactorial analysis showed that advanced age, concurrent chronic diseases, homocysteine >15 mmol/L, lymphocyte to monocyte ratio <3.713, albumin to globulin ratio <1.527, and geriatric nutritional risk index <99.837 were risk factors for poor prognosis in elderly patients with acute ischemic stroke. AUC of the nomogram was 0.774 (95%CI:0.730~0.818); mean absolute error of conformity between the true and predicted value was 0.011 by bootstrap internal validation method; Brier score was 0.152; Hosmer-Lemeshow test P=0.733, and decision curve showed high risk probability value ranged between 0.03 and 0.65. Conclusion The constructed nomogram based on inflammatory and nutritional indicators have high application value and could be used as a tool to predict the poor prognosis of acute ischemic stroke in the elderly.

Key words: ischemic stroke, inflammation, nutrition assessment, model, statistics