Hemorrhagic risk factors after rt PA thrombolysis in acute cerebral infarction.

OBJECTIVE

This study aims to investigate the risk factors of hemorrhagic transformation (HT) after thrombolysis with recombinant tissue plasminogen activator (rt-PA) in patients with acute cerebral infarction (ACI) and establish a logistic regression equation and the risk prediction model.

PATIENTS AND METHODS

One hundred and ninety patients with ACI were divided into the HT group (n=20) and non-HT group (n=170) according to whether HT occurred within 24 hours after rt-PA thrombolysis. The clinical data were collected for analyzing the influencing factors, and a logistic regression analysis model was then established. Besides, patients in the HT group were further grouped into symptomatic hemorrhage (n=7) and non-symptomatic hemorrhage (n=13) according to the type of hemorrhage. The clinical diagnostic value of risk factors in symptomatic hemorrhage after thrombolysis in ACI was analyzed using the ROC curve.

RESULTS

We found that history of atrial fibrillation, time from onset to thrombolysis, pre-thrombolytic glucose, pre-thrombolytic National Institute of Health Stroke Scale (NIHSS) score, 24-hour post-thrombolytic NIHSS score, and proportion of patients with large cerebral infarction were all the influencing factors of HT risk after rt-PA thrombolysis in patients with ACI (p<0.05). Logistic regression analysis model was established with an accuracy of 88.42% (168/190), a sensitivity of 75.00% (15/20), and a specificity of 90.00% (153/170). The time from onset to thrombolysis, pre-thrombolytic glucose, and 24-hour post-thrombolytic NIHSS score had higher clinical value in predicting the risk of HT after rt-PA thrombolysis, with the AUCs of 0.874, 0.815 and 0.881, respectively. Blood glucose and pre-thrombolytic NIHSS score were independent risk factors related to symptomatic hemorrhage after thrombolysis in ACI (p<0.05). The AUCs for predicting symptomatic hemorrhage alone and in combination were 0.813, 0.835, and 0.907, respectively, with sensitivities of 85.70%, 87.50% and 90.00%, and specificities of 62.50%, 60.00%, and 75.42% respectively.

CONCLUSIONS

The establishment of a prediction model based on the risk factors of HT after rt-PA thrombolysis had a good predictive value in patients with ACI. This model was helpful in guiding clinical judgment and improving the safety of intravenous thrombolysis. Early identification of symptomatic bleeding risk factors provided a reference for clinical treatment and prognostic measures of patients with ACI.