European Review for Medical and Pharmacological Sciences 2021; 25: 3066-3073 Delayed intracranial hemorrhage after mild traumatic brain injury in patients on oral anticoagulants: is the juice worth the squeeze? M. COVINO1,2, A. MANNO1, G.M. DELLA PEPA3, A. PICCIONI1, G. TULLO1, M. PETRUCCI1, S. NAVARRA1, F. SARDEO1, E. TORELLI1, R. NICOLÒ1, B. SIMEONI1, L. CARBONE1, S. GAUDINO2,4, F. FRANCESCHI1,2 1Emergency Medicine, Fondazione Policlinico Universitario A. Gemelli, IRCCS, Rome, Italy 2Università Cattolica del Sacro Cuore, Rome, Italy 3Institute of Neurosurgery, Fondazione Policlinico Universitario A. Gemelli, IRCCS, Rome, Italy 4Department of Neuroradiology, Institute of Radiology, Fondazione Policlinico Universitario A. Gemelli, IRCCS, Rome, Italy Abstract. – OBJECTIVE: Mild Traumatic a routine control CT scan seems advisable only Brain Injury (MTBI) in anticoagulated patients is for patients presenting a clinical deterioration. a common challenge for Emergency Department Larger, prospective trials are required to clari- (ED) Physicians. Anticoagulation is considered fy the safety profile of DOACs vs. VKA in MTBI. a risk factor for developing delayed intracranial hemorrhage (ICH) after MTBI. The occurrence of Key Words: this event in patients on Vitamin K Antagonists Mild traumatic brain injury, Anticoagulation, Direct (VKA) or Direct Oral Anticoagulants (DOACs) re- oral anticoagulants, Intracranial hemorrhage. mains unclear. Primary endpoint: to analyze the role of anticoagulants as risk factors for devel- oping delayed ICH after MTBI and evaluate the Introduction indications to repeat a cranial computed tomog- raphy (CT) after a period of observation. Sec- Traumatic brain injury (TBI) represents one ondary endpoint: to assess the difference in the prevalence rate of delayed ICH in patients on of the most common causes of morbidity and VKA versus those on DOACs. mortality worldwide. In Europe, the incidence PATIENTS AND METHODS: We evaluated all rate is 235/100,000, with an average mortality consecutive patients admitted to our ED for MT- rate of 15/100,000 and an estimated annual cost BI, which had a control CT for late ICH after a of 33 billion Euros1. Approximately 80% of all negative CT at admission. We used a propen- TBI are mild (MTBI)2. The opportunity to sub- sity score match (PSM) on factors affecting the need for oral anticoagulation to adjust the com- mit patients with MTBI to computed tomography parison between anticoagulated vs. non-antico- (CT) at admission to the emergency department agulated patients for the baseline clinical char- (ED) is well established2,3. However, there is still acteristics. a lack of consensus on routine CT repeat after a RESULTS: Among 685 patients enrolled, 15 period of clinical observation3,4. MTBI common- (2.2%) developed ICH at control CT. After PSM, ly occurs in older patients, which have increased the incidence of ICH, although slightly high- fall risk, and commonly need anticoagulation for er, was not statistically different in anticoagu- 5 lated patients vs. non-anticoagulated (2.3% vs. several clinical conditions . Assumption of long- 0.6%, p=0.371). Among the 111 patients on VKA, term oral anticoagulants is traditionally consid- 5 (4.5%) had a late ICH, compared to 4 out of 99 ered an independent risk factor for developing (4.0%) on DOACs; the difference was not statis- both immediate than delayed intracranial hemor- tically significant (p=0.868). rhage (ICH) after MTBI6 but not all investigators CONCLUSIONS: The risk of developing de- agree7. Moreover, while for patients on vitamin layed ICH after MTBI in patients on anticoagu- K antagonist (VKA), available data are relatively lation therapy is low. After correction for base- 8,9 line covariates, the risk does not appear higher consistent , minor evidence exists for patients on compared to non-anticoagulated patients. Thus, Direct Oral Anticoagulants (DOAC)10-12. Howev- 3066 Corresponding Author: Marcello Covino, MD; e-mail: [email protected] Delayed intracranial hemorrhage and anticoagulation in MTB er, current studies suggest that the risk of adverse • Patients with a known history of inherited co- outcomes following a MTBI could be low for agulation disease; these patients13. • Patients that had a positive finding at first CT This study’s primary endpoint is to analyze the scan assessment. role of anticoagulants as risk factors for develop- ing delayed ICH after MTBI. The secondary end- For all patients included, we extracted the fol- point is to assess the difference in the prevalence lowing data by manually reviewing the clinical rate of delayed ICH in patients on VKA versus records: those on DOACs. • Demographics, including sex and age at the index admission; • Anticoagulant therapy at the time of injury, in- Patients and Methods cluding VKA DOACs or low-molecular-weight heparin (either therapeutic or prophylaxis dos- Study Design and Setting ing); The present paper reports the result of an ob- • Antiplatelet drug therapy, including aspirin or servational, cross-sectional study conducted in an clopidogrel; Emergency Department admitting about 77,000 • Presence of comorbidities, including history patients per year, in an urban teaching hospital. of neoplasia, neurodegenerative diseases, cere- The study was conducted over a three-year peri- brovascular diseases, thrombocytopenia, epi- od between January 1st, 2016 and December 31st, lepsy, and alcohol abuse; 2018. Our institution serves a metropolitan area, • Neurological and physical examination data and our hospital is a referral center for several at admission and during the ED observation peripheral Emergency Departments, and it is a period. major trauma center. CT Scan Evaluation Patients Selection and Data Extraction Decision-making to perform CCT was always We performed an automatized search of the taken by a board-certified emergency physician ED electronic clinical records based on admis- based on clinical evaluation and NICE guide- sion and discharge diagnosis. We included all lines4. Axial CT images were acquired at 2.5 mm consecutive records of patients admitted to our slices on a 64 slide CT scan [Revolution CT, GE ED for MTBI as chief complaint who, after a first Healthcare]. Cranial computed tomography scan negative CT scan at admission, repeated a CT interpretations were performed in all cases by ex- scan 24 hours later. MTBI was defined as TBI perienced neuro-radiologists. CT was considered with Glasgow Coma Score (GCS) ≥ 13, loss of positive if any kind of acute intracranial bleeding consciousness < 30 minutes, and post-traumatic was found (regardless of the amount), including amnesia < 24 hours [3]. subarachnoid hemorrhage, subdural hematoma, Our standard ED protocol indicates for all MT- epidural hematoma, intra-parenchymal hemor- BI patient a 6-hour observation. Patients receive a rhage, and cerebral contusion. Results of the CT head CT scan at admission, based on emergency scans were obtained from radiologists’ reports. physician evaluation. Patients who experience any clinical worsening during the observation Event Adjudication period [episode of epilepsy, vomit ≥ 2 episodes, We evaluated the number of patients that had a persistence of GCS < 15, prolonged amnesia, positive control CT scan, after an index negative persistent headache], are prescribed a prolonged CT assessment in our ED. observation and a 24-hour repeat CT scan. All patients on anticoagulant therapy (either VKA or Statistical Analysis DOAC) get a prolonged observation and a con- Continuous variables are reported as median trol CT scan at 24 hours from the index control. [interquartile range]. Categorical variables are Control CT scan could be anticipated based on reported as absolute numbers (%). Statistical uni- evolving clinical findings. variate comparison with respect of primary study We excluded from analysis: endpoint was assessed by Mann-Whitney U test • Trauma not classified as MTBI; for continuous variables, and Chi-square test • Patients <18 years of age; (with Yates correction or Fisher test if appropri- • Pregnant women; ate) for categorical variables. 3067 M. Covino, A. Manno, G.M. Della Pepa, A. Piccioni, G. Tullo, et al. Since this is a retrospective study, and criteria excluding criteria, we enrolled 685 patients in the for prescription of anticoagulation therapy in- study. The median age was 79 years [Interquartile clude several comorbidities and are often age-de- range 64.5-86]; 329 (48%) were males (Table I). pendent, we used a propensity score matching Overall, 210 patients (30.7%) were on anticoagu- (PSM) analysis to compare non-anticoagulated lant therapy. Among them, 111 (52.8%) were on vs. anticoagulated patients in order to adjust the VKA and 99 (47.2%) on DOACs. In the DOACs result of our study for these baseline factors. group 23 (11.0%) were on dabigatran, 37 (17.6%) PSM was generated using a logistic regression on apixaban, 31 (14.8%) on rivaroxaban and 9 model on the covariates considered potentially (4.3%) on edoxaban. influencing the decision to prescribe anticoagu- Control CT was positive for delayed ICH in 15 lation and adding any confounding factors iden- (2.2%) cases (Table I). tified as significantly associated to hemorrhage. Of the 15 patients with ICH after MTBI, one Variable considered for PSM were age, gender, refused admission to the ward (no anticoagulant ASA, clopidogrel, LMW heparin, alcohol abuse, therapy assumed), three were discharged from malignancy, and high energy trauma. Additional the ED after a 24-hour clinical control.