Author + information
- Saurav Chatterjee, MD∗ (, )
- Gregory Y.H. Lip, MD and
- Jay Giri, MD, MPH
- ↵∗Cardiovascular Medicine, Mount Sinai St. Luke’s-Roosevelt Hospitals, 1111 Amsterdam Avenue, New York, New York 10025
Intracranial hemorrhage (ICH) is one of the dreaded complications of thrombolytic therapy for acute pulmonary embolism (PE) (1). Prior registries such as the RIETE registry (Computerized Registry of Patients with Venous Thromboembolism) and the International Cooperative Pulmonary Embolism Registry (ICOPER) have noted risk of major bleeding including ICH at 2.4% and 3%, respectively (1), indicating need for identifying the same accurately. We identified patients with PE treated with thrombolytic therapy who may be at relatively high risk of ICH from an administrative database, and compared the ability of 2 commonly used bleeding risk scores used for risk prediction in anticoagulation-treated patients to accurately predict ICH.
We used data from patients discharged from short-stay hospitals in the United States from 2011 to 2012 with PE (ICD-9 diagnosis codes 415.1, 415.11, 415.13, and 415.19) who received thrombolytic therapy (ICD-9 procedure code 99.10) and the proportion with ICH from the Nationwide Inpatient Sample (Healthcare Cost and Utilization Project, Agency for Healthcare Research and Quality), which includes 20% of all hospital discharges in the United States (2). Pre-existing medical conditions were identified using the Deyo modification of the Charlson comorbidity index (3). Two commonly used bleeding risk scores used for assessing bleeding risk in patients on anticoagulation—the HASBLED (4), and ATRIA (5)—were assessed to identify ability to accurately identify risk of ICH in PE patients receiving thrombolytic therapy. The individual components of the HAS-BLED score were identified from administrative codes: hypertension (ICD-9 codes 401, 405), abnormal renal function (codes 585), abnormal liver function (codes 571, 573.9), stroke (438), bleeding (459), coagulopathy (286.9), elderly (age >65 years), and alcohol (303.9)/drug use (304.9). Similar methodology was used to identify components of the ATRIA score: anemia (285.9), end-stage renal disease (585.6), age >65 years, prior hemorrhage (998.11) and hypertension (401, 405). ICH events were identified using codes 430 to 432. For the purpose of simplicity in risk assessment, the variables were dichotomized. Standard receiver-operating characteristic curves were constructed to evaluate the ability of the HAS-BLED and the ATRIA risk scores to predict ICH in PE patients receiving thrombolytic agents. Model calibration was assessed with multiple bootstrapping with replacement.
From 2011 to 2012, 2,727 patients received thrombolytic therapy for PE; the prevalence of intracerebral hemorrhage was 48 (1.8%). Patients who developed ICH tended to be older, had a higher prevalence of prior history of cerebrovascular disorders and paralysis, and had a longer length of hospital stay (Table 1), although other comorbidities were comparable. Both the HASBLED and the ATRIA scores predicted in-hospital ICH risk (p < 0.001 for both), (receiver-operating characteristic: 0.57 [95% confidence interval: 0.55 to 0.59] and 0.53 [95% confidence interval: 0.51 to 0.55], respectively), but cannot be considered adequate for risk prediction, even if statistically significant. There was no significant difference in discrimination between the scores.
Currently available bleeding scores are only able to moderately predict the risk of ICH after thrombolytic agent use in PE, thus identifying an urgent need for more precise risk prediction tools, acknowledging limitations of using an administrative dataset, and dichotomizing predictor variables for deriving the same.
Please note: Dr. Chatterjee had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Dr. Lip is a consultant for Bayer/Jensen J&J, Astellas, Merck, Sanofi, BMS/Pfizer, Biotronik, Medtronic, Portola, Boehringer Ingelheim, Microlife, and Daiichi-Sankyo; and is a speaker for Bayer, BMS/Pfizer, Medtronic, Boehringer Ingelheim, Microlife, Roche, and Daiichi-Sankyo. Dr. Giri has reported that he has no relationships relevant to the contents of this paper to disclose. P.K. Shah, MD, served as Guest Editor for this paper.
- American College of Cardiology Foundation
- Healthcare Cost and Utilization Project (HCUP)
- Fang M.C.,
- Go A.S.,
- Chang Y.,
- et al.