Factors influencing hemorrhagic transformation after mechanical thrombectomy for acute anterior circulation large vessel occlusion stroke and prediction modeling

Introduction
The aim was to explore the factors influencing hemorrhagic transformation (HT) after mechanical thrombectomy (MT) in acute anterior circulation large vessel occlusion stroke (ALVOS) and establish a corresponding prediction model.

Material and methods
A retrospective study was conducted on 180 ALVOS patients who underwent MT in our hospital between May 2022 and December 2023. The patients were divided into a bleeding group (134 cases) and a non-bleeding group (46 cases) based on whether there was intracranial hemorrhage in the immediate follow-up head CT after surgery. Logistic regression analysis was performed to explore the factors influencing HT after MT in ALVOS patients. A logistic regression prediction model based on risk factors was constructed. The predictive ability of the model was validated using receiver operating characteristic (ROC) curves.

Results
Significant differences were detected between the bleeding group and the non-bleeding group in terms of age, diabetes, NIHSS score on admission, time from onset to admission, time from onset to vascular recanalization, number of embolectomy attempts, and application of tirofiban (p < 0.05). Diabetes, NIHSS score on admission, time from onset to admission, time from onset to vascular recanalization, number of embolectomy attempts, and application of tirofiban were all significant factors influencing HT in ALVOS patients who underwent MT (p < 0.05). The logistic risk prediction model was constructed using the following model: Logit (P) = 0.625 × combined diabetes + 0.071 × NIHSS score + 0.035 × onset to hospital time + 2.321 × onset to vascular recanalization time + 1.461 × number of embolectomy attempts + 0.993 × application of tirofiban. The model had the likelihood ratio chi square (c2) = 196.85, DF = 6, p < 0.001. The model demonstrated good fit, with c2 = 7.687, DF = 6 and p = 0.724. ROC curves based on predicted and true values showed that when the logistic score > 2.03, the AUC was 0.882, 95% CI was 0.792-0.175, Z = 20.331, p < 0.001, with the predictive sensitivity of 85.35% and the specificity of 85.74%.

Conclusions
Age, diabetes, NIHSS score on admission, time from onset to admission, time from onset to vascular recanalization, number of thrombectomy attempts, and use of tirofiban were independent risk factors for HT after MT in ALVOS patients. A logistic risk prediction model incorporating independent risk factors had some predictive value for HT after thrombolysis.