Predicting cerebral infarction in COPD patients: an individualized nomogram based on arterial oxygen saturation

To investigate the correlation between arterial oxygen saturation (SaO₂) and the occurrence of cerebral infarction (CI) in patients with chronic obstructive pulmonary disease (COPD), and to develop a nomogram model based on SaO₂ to predict the probability of CI in COPD patients.

This retrospective study analyzed the clinical data of 846 COPD patients admitted to the Affiliated Hospital of Guangdong Medical University from June 2018 to December 2019. Logistic regression analysis was used to identify risk factors for CI, and the Wald chi-square test was applied to select predictors for inclusion in the nomogram. The performance of the model was evaluated and internally validated. External validation was conducted using data from 290 COPD patients prospectively enrolled at the Second Affiliated Hospital of Guangdong Medical University between January and October 2024.

A total of 846 COPD patients were included, with 592 assigned to the training cohort and 254 to the internal validation cohort. Predictive factors incorporated into the nomogram included systemic immune-inflammation index (SII), age, hypertension, cardiovascular disease (CVD), paraplegia, and SaO₂. The nomogram demonstrated strong predictive accuracy and calibration, with AUCs of 0.85 (95% CI: 0.82-0.89) in the training cohort, 0.89 (95% CI: 0.85-0.94) in the internal validation cohort, and 0.90 (95% CI: 0.86-0.94) in the external validation cohort.

COPD is prone to cerebral infarction, and we verified the relationship between COPD and cerebral infarction through a visualogram model. The incidence of cerebral infarction in COPD patients is affected by systemic immune-inflammatory index (SII), age, hypertension, cardiovascular disease (C), hemiplegia, and SaO₂, and the nomogram model for risk prediction based on SaO₂ has good predictive efficacy, which can provide a reference forbral infarction in COPD patients in clinical practice. However, this study was retrospective, the sample size of the different subtypes of ischemic stroke was small, the risk analysis of each subtype of ischemic stroke could not be performed, and there may be some unmeasured confounding factors, which had a certain impact on the of the study. In the future, we need multi-center prospective studies to verify the effectiveness and practicality of the nomogram, and the basic mechanism of each risk factor also to be studied.