# Construction and validation of a nomogram for predicting 3-month outcome in elderly patients with nonvalvular atrial fibrillation-induced acute ischemic stroke

## TL;DR

This study creates a tool to predict 3-month outcomes in elderly patients with a specific type of stroke caused by atrial fibrillation, using factors like medical history and lab results.

## Contribution

A novel nomogram was developed and validated for predicting outcomes in elderly patients with nonvalvular atrial fibrillation-induced acute ischemic stroke.

## Key findings

- Stroke history, NIHSS score, HbA1c, D-dimer, NLR, and LAD were identified as independent risk factors for poor outcomes.
- The nomogram demonstrated excellent discriminative ability with an AUC of 0.933 before and after validation.
- Calibration and decision curve analysis confirmed the nomogram's clinical applicability.

## Abstract

Cardiogenic embolism caused by atrial fibrillation (AF) leads to higher disability, mortality, and recurrence rates compared to stroke independent of AF, resulting in a poorer prognosis for patients. Effective risk assessment and timely clinical intervention are essential. This study aimed to develop and validate a personalized nomogram to predict the 3-month outcomes for elderly patients with nonvalvular atrial fibrillation (NVAF) induced acute ischemic stroke (AIS).

A retrospective cohort study was implemented at Hefei First People’s Hospital. Participants were patients diagnosed with NVAF-induced acute ischemic stroke (NVAF-AIS) who fulfilled the study’s inclusion criteria. Data collection encompassed baseline demographic, clinical, and laboratory information. The primary endpoint was the 3-month outcome, evaluated using the modified Rankin Scale (mRS). To identify potential predictors, univariate logistic regression and the least absolute shrinkage and selection operator (LASSO) regression algorithm were employed. Subsequently, a binary regression model was established, and internal validation was conducted using bootstrap resampling with 1,000 iterations. The assessment tools included receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Ultimately, a nomogram was constructed to forecast the 3-month outcomes for this demographic.

A total of 178 patients were included, of whom 95 (53.3%) had a poor outcome (mRS > 2). Independent risk factors for poor outcomes in NVAF-AIS patients included stroke history (OR = 9.140; 95% CI: 3.481–26.923), NIHSS score (OR = 1.167; 95% CI: 1.071–1.284), glycated hemoglobin (HbA1c) (OR = 2.211; 95% CI:1.573–3.220), D-dimer (OR = 1.157; 95% CI: 1.022–1.361), neutrophil-to-lymphocyte ratio (NLR) (OR = 1.531; 95%CI:1.242–1.972), and left atrial diameter (LAD) (OR = 1.163; 95%CI: 1.072–1.280). A nomogram was created based on these factors. The area under the ROC curve (AUC) for the nomogram was 0.933 (95%CI:0.897–0.969) before and 0.933(95%CI:0.895–0.964) after internal validation, demonstrating good discriminative ability. The nomogram also showed excellent calibration and clinical applicability, as confirmed by calibration curve analysis and DCA.

Stroke history, NIHSS score, HbA1c, D-dimer, NLR, and LAD are independent risk factors for poor outcomes in elderly patients with NVAF-AIS. The nomogram, integrating these factors, provides intuitive, individualized predictions for the risk of poor outcomes, aiding in the selection of treatment options for these patients.