Article, Neurology

Incidence and risk factors of delayed intracranial hemorrhage in the emergency department

a b s t r a c t

Objectives: This study was performed to identify the risk factors for delayed intracranial hemorrhage and develop a risk stratification system for disposition of head trauma patients with negative initial brain imaging.

Methods: The data source was National Health Insurance Service-National Sample Cohort of Korea. We analyzed adult patients presenting to the ER from January 2004 to September 2012, who underwent brain imaging and discharged with or without short-term observation no longer than two days. The primary outcome was defined as any Intracranial bleeding within a month defined by a new appearance of any of the Diagnostic codes for in- tracranial hemorrhage accompanied by a new claim for brain imaging(s) within a month of the index visit. We performed a multivariable logistic regression analysis and built a parsimonious model for variable selection to de- velop a simple scoring system for risk stratification.

Results: During the study period, a total of 19,723 head injury cases were identified from the cohort and a total of

149 cases were identified as having delayed intracranial hemorrhage within 30 days. In multivariable logistic re- gression model, old age, craniofacial fracture, neck injury, diabetes mellitus and hypertension were independent risk factors for delayed intracranial hemorrhage. We constructed the parsimonious model included age, cranio- facial fracture and diabetes mellitus. The score showed area under the curve of 0.704 and positive predictive value of the score system was 0.014 when the score >= 2.

Conclusions: We found old age, associated craniofacial fracture, any neck injury, diabetes mellitus and hyperten- sion are the independent risk factors of delayed intracranial hemorrhage.

(C) 2017

Introduction

Head trauma is one of the most common causes of visits to emergen- cy department (ED) and intracranial hemorrhage is one of its most se- vere complications [1,2]. Brain imaging including brain computed tomography (CT) or magnetic resonance imaging (MRI) can effectively rule out immediate intracranial hemorrhages in most cases, however, delayed intracranial hemorrhages have been reported to occur after a negative test which can result in significant complication [3,4]. Current- ly, it is recommended that patients with high risk of intracranial bleed- ing such as those with anticoagulation should undergo an initial CT scan and be admit for overnight observation during the first 24 h, and receive

* Corresponding author.

E-mail address: [email protected] (J. Kim).

1 Byunghyun Kim and Hyeonjeong Jeong contributed equally to this work.

a follow-up CT scan before discharge [5]. However, current knowledge about its risk factors have been limited because previous studies tried to identify them in Anticoagulated patients with rather small population size. In addition, there has been no risk stratification system previously developed [6,7].

It would be almost impossible as well as impractical to admit every ED patient with head trauma after a negative brain imaging considering the rare incidence of delayed intracranial hemorrhage [8,9]. Instead, a practical approach would be to risk stratify the patients with head trau- ma and provide them rational disposition plans based on the risks [10, 11]. For more comprehensive assessment of the risk for delayed intra- cranial hemorrhage, further studies to identify its general risk factors would be critical.

In this study, we conducted a population-based observational study using a nationally representative cohort to identify the risk factors for delayed intracranial hemorrhage and develop a risk stratification

http://dx.doi.org/10.1016/j.ajem.2017.08.009 0735-6757/(C) 2017

Table 1

Baseline characteristics of the study population.

No delayed hemorrhage within a month

Delayed pa hemorrhage within

a month

Table 3

Multivariable logistic regression model for the development of a new intracranial hemor- rhage within the first 30 days.

(N = 19,574) (N = 149)

Age, years b0.001

b30 5002 (25.6%) 12 (8.1%)

30-59 10,338 (52.8%) 66 (44.3%)

>= 60 4234 (21.6%) 71 (47.7%)

Sex, male 12,210 (62.4%) 103 (69.1%) 0.107

Syncope-related 596 (3.0%) 8 (5.4%) 0.161

Associated injuries

Laceration

8079 (41.3%)

52 (34.9%)

0.136

Craniofacial fracture

4718 (24.1%)

48 (32.2%)

0.027

Tooth injury

226 (1.2%)

0 (0.0%)

0.351

Neck injury

3415 (17.4%)

37 (24.8%)

0.024

With C-spine injury

3307 (16.9%)

34 (22.8%)

0.070

Without C-spine injury

108 (0.6%)

3 (2.0%)

0.068

Oral antithrombotic agents

1456 (7.4%)

23 (15.4%)

b0.001

Aspirin

1193 (6.1%)

21 (14.1%)

b0.001

Clopidogrel

352 (1.8%)

6 (4.0%)

0.085

Vitamin-K antagonist

Comorbidities

75 (0.4%)

1 (0.7%)

1.000

Atrial fibrillation or flutter

130 (0.7%)

2 (1.3%)

0.612

Diabetes mellitus

1376 (7.0%)

31 (20.8%)

b0.001

Hypertension

3621 (18.5%)

62 (41.6%)

b0.001

Dyslipidemia

3098 (15.8%)

39 (26.2%)

0.001

Chronic renal failure

172 (0.9%)

3 (2.0%)

0.302

End-stage renal disease

39 (0.2%)

1 (0.7%)

0.718

Ischemic stroke

491 (2.5%)

11 (7.4%)

b0.001

Ischemic heart disease

889 (4.5%)

20 (13.4%)

b0.001

Heart failure

602 (3.1%)

14 (9.4%)

b0.001

Advanced liver disease

262 (1.3%)

4 (2.7%)

0.288

Malignancy

483 (2.5%)

8 (5.4%)

0.045

Dementia

347 (1.8%)

5 (3.4%)

0.253

Types of delayed hemorrhage (within a month)

Subdural 72 (48.3%)

Subarachnoid 32 (21.5%)

Intracerebral 24 (16.1%)

Extradural 19 (12.8%)

Others 2 (1.3%)

Diagnosis within a week 114 (76.5%)

a Wilcoxon rank sum, chi-square, or Fisher’s exact test was performed as appropriate.

system for rational disposition of the patients with head injury with negative initial brain imaging.

Materials and methods

Study design and setting

The data source was National Health Insurance Service-National Sample Cohort (NHIS-NSC), a population-based cohort established by the Korean NHIS [12]. It contains claim information of one million indi- viduals who were randomly sampled after stratification from almost en- tire Korean population. It provides diagnostic codes based on international classification of diseases (ICD)-10 Coding system,

The model AUC was 0.727 (0.687-0.767).

Odds ratio

p

Age 30-59

2.51 (1.40-4.92)

0.004

Age >= 60

4.92 (2.57-10.09)

b0.001

Oral antithrombotic agents

0.71 (0.41-1.20)

0.214

CranioFacial fractures

1.99 (1.39-2.82)

b0.001

neck injuries

1.60 (1.08-2.31)

0.015

Diabetes mellitus

1.87 (1.17-2.92)

0.007

Hypertension

1.58 (1.01-2.45)

0.043

Dyslipidemia

0.84 (0.54-1.29)

0.439

Ischemic stroke

1.49 (0.73-2.75)

0.233

Ischemic heart disease

1.59 (0.89-2.73)

0.104

Heart failure

1.39 (0.73-2.49)

0.292

Malignancy

1.21 (0.54-2.38)

0.604

prescription and procedure codes and related costs as well as demo- graphic information such as age, sex and socioeconomic status. It also has information about disability and death based on national disability registration data and death certificates, respectively. We used its most recent release which contains claim data from 2002 to 2013. The de- tailed descriptions of the cohort data can be found in a previous paper [12].

Selection of cases, outcome measures and related definitions

Inclusion criteria was ED head trauma cases of adult (aged >= 15) pa- tients with ICD codes for head injury (S00x-S09x) who underwent brain imaging (CT or MRI) and discharged with or without short-term observation no longer than two days from Jan. 2004 to Sep. 2013. Cases with any of the diagnostic codes for Intracranial injury other than concussion (S06.1x-S06.9x) or accompanying non-traumatic in- tracranial hemorrhage (I60.x, I61.x, I62.x) were excluded. Cases with previous entry of diagnostic codes for traumatic or non-traumatic intra- cranial bleeding (S06.24x, S06.34x, S06.35x, S06.4x-S06.6x, I60.x-I62.x) within two years prior to the index visit were also excluded because a new entry of the diagnostic codes for intracranial hemorrhage in the population could be referring to the previous hemorrhage events.

The primary outcome was defined as any delayed intracranial bleed- ing within a month of the index visit to ER. Secondary outcomes were the same events occurring within a week and 3 months after the index visit to ER. We defined a delayed intracranial hemorrhage event by a new appearance of a combination of 1) new diagnostic codes for in- tracranial hemorrhage and 2) a new procedure code for brain imaging (CT or MRI). If the claim record for the index visit and the claim records with the presumed delayed intracranial hemorrhage were issued at the same day, we assumed that the one with a diagnosis of intracranial hemorrhage are occurring later than the others.

Comorbidities, use of antithrombotic agents and associated injuries at the time of the index visit were also defined based on diagnostic codes and procedure codes. Significant complications including neuro- surgery and death due to head injury following the second visit for de- layed hemorrhage were also assessed. The detailed descriptions of the

Table 2

Types of delayed intracranial hemorrhagic events observed during the first 90 days and the incidence of major complications (neurosurgery and death by head injury).

0-7 days (N = 114) 8-30 days (N = 35) 31-90 days (N = 50)

Total (%)

Neurosurgery

Death

Any

Total (%)

Neurosurgery

Death

Any

Total (%)

Neurosurgery

Death

Any

SDH

54 (47.4)

6

1

6

18 (51.4)

7

2

9

28 (56)

6

1

6

ICH

17 (14.9)

2

0

2

7 (20.0)

1

0

1

14 (28)

2

0

2

EDH

16 (14.0)

0

0

0

3 (8.6)

0

0

0

3 (6)

0

0

0

SAH

26 (22.8)

0

1

1

6 (17.1)

0

0

0

2 (4)

0

1

1

Others

1 (0.9)

0

0

0

1 (2.9)

1

0

1

3 (6)

0

0

0

Note: “death” indicate death due to head injury (S00-S09) and “any” indicates the number of patients with any of the complication (neurosurgery or death).

Table 4

Parsimonious model for the development of a new intracranial hemorrhage within the

first 30 days.

Odds ratio p

Age 30-59 2.69 (1.51-5.26) 0.002

Age >= 60 6.54 (3.59-12.91) b0.001

Craniofacial fracture 1.91 (1.33-2.69) b0.001

Diabetes mellitus 2.12 (1.37-3.20) b0.001

The model AUC was 0.705 (0.665-0.745) and the beta coefficients were 0.99, 1.88, 0.64

and 0.75, respectively.

definitions used in this study are available in the Supplementary Tables 1, 2 and 3.

Statistical analysis

Categorical variables were reported using frequencies and propor- tions, while continuous variables were reported using medians and in- terquartile ranges (IQRs). The age variable was binned using supervised discretization [13]. Wilcoxon’s rank-sum test, the chi- square test, or Fisher’s exact test was performed, as appropriate, for comparisons between groups. We constructed a multivariable logistic regression model using variables whose p-value for the difference be- tween the groups was b 0.1 to identify the independent risk factors for the development of any delayed intracranial hemorrhage within a month. We also built a parsimonious model using Bayiesian information criterion (BIC) for variable selection to develop a simple scoring system

for risk stratification. The goodness of fit of the models was assessed with the Hosmer-Lemeshow test. Calibration of the scoring system was tested by checking the incidence of the event in each stratum of the score ranges. The results of the logistic regression analyses were presented as odd ratios (ORs), and their 95% confidence intervals (CIs) were also presented. p-Values b 0.05 were considered significant. All data handling and statistical analyses were performed using R- packages version 3.3.2 (R Foundation for Statistical Computing, Vienna, Austria).

Results

During the study period, a total of 19,723 head injury cases were identified from the cohort (Table 1). A total of 149 cases were identified as having delayed intracranial hemorrhage within 30 days (7.6 cases per 1000 visits). Delayed intracranial hemorrhage was more prevalent in patients with older age (p b 0.001), craniofacial fracture (p = 0.027), neck injury (p = 0.024), and the use of antithrombotic agents (p b 0.001). Comorbidities associated with increased chance of having a de- layed hemorrhage event were diabetes mellitus (p b 0.001), hyperten- sion (p b 0.001), dyslipidemia (p = 0.001), ischemic stroke (p b 0.001), ischemic heart disease (p b 0.001), heart failure (p b 0.001) and malignancy (p = 0.045). Table 2 provides the detailed descriptions of the types of the first hemorrhages and their complications during the first three months (0-7 days, 8-30 days and 31-90 days) after the index visit to ED. SDH was the dominant type across the three periods. The most common type of hemorrhage diagnosed during the first month

Fig. 1. Incidence of delayed intracranial hemorrhage per 1000 cases stratified by score system.

Table 5

Test characteristics of delayed hemorrhage scoring system using different cut-off values.

Sensitivity

Specificity

PPV

NPV

N

Score >= 1

0.97 (0.92-0.99)

0.17 (0.17-0.18)

0.009 (0.007-0.010)

0.999 (0.997-1.000)

16,357 (82.9%)

Score >= 2

0.70 (0.62-0.77)

0.64 (0.63-0.64)

0.014 (0.012-0.017)

0.996 (0.995-0.997)

7212 (36.6%)

Score >= 3

0.50 (0.41-0.58)

0.78 (0.77-0.78)

0.017 (0.013-0.021)

0.995 (0.994-0.996)

4408 (22.3%)

Score >= 4

0.22 (0.16-0.30)

0.93 (0.93-0.93)

0.024 (0.016-0.033)

0.994 (0.992-0.995)

1382 (7.0%)

PPV, positive predictive value; NPV, negative predictive value.

was subdural hemorrhage (SDH, n = 72/149, 48.3%) and most of the de- layed hemorrhages were diagnosed within the first week (n = 114/149, 76.5%). Neurosurgery was performed in 8/114 (7.0%), 9/35 (25.7%) and 19/50 (38.0%) patients, respectively, while the death due to head injury was followed in 2/114 (1.8%), 2/35 (5.7%) and 2/50 (4.0%) patients, respectively.

A multivariable logistic regression model was constructed using the variables with their p-value for difference between the groups b 0.1 (Table 3). Old age (OR = 2.51, 95% CI 1.40-4.92, p = 0.004 for age 30-59, and OR = 4.92, 95% CI 2.57-10.09, p b 0.001 for age >= 60 relative to age b 30), craniofacial fracture (OR = 1.99, 95% CI 1.39-2.82, p b 0.001), neck injury (OR = 1.60, 95% CI 1.08-2.31, p = 0.015), diabetes mellitus (OR = 1.87, 95% CI 1.17-2.92, p = 0.007) and hypertension (OR = 1.58, 95% CI 1.01-2.45, p = 0.043) were independent risk factors for delayed intracranial hemorrhage. We constructed a parsimonious logistic regression model using BIC for variable selection for develop- ment of a simple risk stratification tool (Table 4). The model included age (OR = 2.69, 95% CI 1.51-5.26, p = 0.002 for age 30-59, and OR

= 6.54, 95% CI 3.59-12.91, p b 0.001 for age >= 60 relative to age b 30), craniofacial fracture (OR = 1.91, 95% CI 1.33-2.69, p b 0.001) and diabe- tes mellitus (OR = 2.12, 95% CI 1.37-3.20, p b 0.001). Based on the par- simonious model, we constructed a scoring system allocating 1.5 for age 30-59, 3.0 for age 60- and 1.0 for both craniofacial fracture and diabetes mellitus. The score showed area under the curve (AUC) of 0.704 (95% CI 0.664-0.744) and showed significant correlation with the incidence of delayed hemorrhage as shown in the calibration test (Fig. 1). Test char- acteristics of the scoring system showed the risk of having a delayed hemorrhage exceeds the 1% margin (positive predictive value = 0.014, 95% CI 0.012-0.017) when the score is >= 2 (Table 5).

We conducted a post-hoc analysis to estimate the age-adjusted inci- dence of delayed intracranial hemorrhage in the users of anti-platelet agents (Table 6). The age-adjusted incidence was 7.4 per 1000 (95% CI 3.8-674.0), which was not significantly different from that for non- users 6.9 per 1000 (95% CI 5.7-8.2).

Discussion

This study is the first population-based longitudinal study of ED pa- tients with head injury conducted to identify the risk factors for delayed intracranial hemorrhage. We found the incidence of delayed intracrani- al hemorrhage within a month is 7.6 per 1000 after a negative brain im- aging and the most dominant type was SDH throughout the first 90 days. Older age, craniofacial fracture, neck injury, diabetes mellitus and hypertension were independent risk factors for development of de- layed intracranial hemorrhage. Delayed intracranial hemorrhage was indeed more common in antithrombotic agent users, however, the dif- ference was mostly due to age differences rather than its independent

effect. The scoring system based on the factors identified from the par- simonious model was able to stratify the risk in general population.

There had been several scientific reports on the risk of delayed intra- cranial hemorrhage in elderly patients having anti-thrombotic agents (Table 7) [14-18]. Although in most studies the incidence of delayed in- tracranial hemorrhage was b 3%, but mortality rate was ranging 0 to 50% with increasing tendency in elderly patients. These findings had con- firmed that delayed intracranial hemorrhage can be a significant prob- lem in the elderly patients with anti-thrombotic prescriptions. In the recent systematized review, Miller et al. reported that the incidence of delayed intracranial hemorrhage as ranging from 5.8 to 72 per 1000 pa- tients [6]. In the study, pooling of data was not attempted because of the heterogeneity of the included studies. In our study, the incidence was

7.6 per 1000 patients. However, the patients included in this study were much younger. If the population is subsetted to the patients aged >= 60, 16.5 patients per 1000 (n = 71/4305, 95% CI 13.0-20.9%) were expected to have delayed intracranial hemorrhage and if the pop- ulation is further subsetted to the patients who were on antithrombotic agents, the incidence becomes 15.9 per 1000 (n = 17/1070, 95% CI 9.6- 25.9%). Both of the estimates were seemed reasonable considering the previously reported incidences of delayed intracranial hemorrhage. However, we think further large, longitudinal prospective studies are required for more accurate estimate of its incidence.

In our study, delayed intracranial hemorrhages was associated with many conditions including old age, craniofacial fracture, neck injury, an- tithrombotic medications and many comorbidities. However, only old age, associated injuries and some of the comorbidities including diabe- tes mellitus and hypertension were independently associated with the risk of delayed intracranial hemorrhage. Among them, old age was the most powerful risk factor for the complication. We think this strong as- sociation with old age was the primary reason for why there was differ- ence in crude incidence of delayed intracranial hemorrhage between users of antithrombotic agents and non-users (Table 1). Indeed, the use of antithrombotic agent risk was not a significant risk factor. This is supported by the findings in the post-hoc analysis where age- adjusted incidence of the delayed hemorrhage event was not signifi- cantly different from that of non-users. However, we still think the use of antithrombotic agents should be always considered in determining the disposition of the patients with head injury because the patients on the medications may have more prolonged and severe hemorrhages once the bleeding begins [11].

The independent associations with the accompanying injuries and comorbidities were some of the novel findings of this study. There has been a knowledge gap in regard to what are the general risk factors for delayed intracranial hemorrhage other than age and anticoagulation. This knowledge gap seems to be due to small number of delayed intra- cranial hemorrhage cases in previous studies, ranging 0 to 11, which

Table 6

Age-adjusted incidence of delayed hemorrhage in patients medicated with anti-platelet agents.a

Crude incidence

Age-adjusted incidence

Incidence in non-medicated

Any anti-platelet agent

16.1 (10.2-24.2)

7.4 (3.8-674.0)

6.9 (5.7-8.2)

Aspirin

17.3 (10.7-26.4)

8.2 (4.1-666.8)

6.9 (5.8-8.2)

Clopidogrel

16.8 (6.2-36.5)

5.7 (2.0-1260.4)

7.4 (6.2-8.7)

a Vitamin-K antagonist was not analyzed because there was only one case of delayed intracranial hemorrhage among patients with the prescription.

Table 7

Risk of delayed intracranial hemorrhage in anticoagulated patients with mild head injuries.

Author

Year

Study design

Intracranial hemorrhage/sample

%

Mortality

95% CI

Kaen [14]

2010

Prospective, 1-center

2/137

1.5

0/2

0.2-5.2

Peck [15]

2011

Retrospective, 1-center

4/362

1.1

0/4

0.3-2.8

Menditto [17]

2012

Prospective, 1-center

7/87

8.0

0/7

3.3-15.9

Nishijima [16]

2012

Prospective, 6-center

4/687

0.6

2/4

0.2-1.5

Schoonman [8]

2014

Retrospective, 2-center

5/211

2.4

1/5

0.7-5.4

McCammack [18]

2015

Retrospective, 1-center

1/144

0.7

0/1

0.1-3.8

Swap [10]

2016

Retrospective, 1-center

11/443

2.5

4/11

1.4-4.4

Subtotal

34/2017

1.7

7/34

1.2-2.3

makes it difficult to perform any multivariable analysis [19]. The associ- ation with accompanying injuries including craniofacial fractures or neck injuries may indicate more severe mechanical force was conducted to brain at the time injury as shown in previous study in which the au- thors argued that facial fractures could be markers of increased risk in brain injury [20]. The reason why the comorbidities including hyperten- sion and diabetes mellitus were associated with the outcomes are not clear. Generally, hypertension is known risk factor of spontaneous intra- cranial hemorrhage and acute lowering of systolic blood pressure in hemorrhage patients is associated with outcome improvement [21]. Also, in recent meta-analysis, diabetes mellitus had modest associations with Spontaneous intracranial hemorrhage [22]. In minor head trauma, we could suggest that vulnerable brain tissues without definite hemor- rhage at initial CT might be more easily to bleed in hypertensive and hy- perglycemic patients and appeared at secondary CT as delayed hemorrhage.

The risk stratification system developed in this study showed fair performances with its AUC of 0.704. If we set its cutoff point as >= 2 where the chance of having delayed intracranial hemorrhage within a month exceeds the 1% margin, the sensitivity and specificity become

0.70 (95% CI 0.62-0.77) and 0.64 (95% CI 0.63-0.64), respectively,

which means about 70% of patients who will develop delayed intracra- nial hemorrhage will be screened initially. However, admitting the pa- tients for close observation seems almost impossible as well as impractical because the number of patients required to be admitted were 4408 (22.3%). Therefore, when determining the disposition of a patient, we think every other aspect of care that can affect the future outcomes should also be considered such as the use of antithrombotic agent which can lead to more severe bleeding and socioeconomic fac- tors that can influence the quality of follow-up cares.

This study has several limitations. First, the study is based on claim

data in which the accuracy of diagnostic codes is limited and significant over- or under-coding could be exist. Specifically, because we defined delayed hemorrhage as a new appearance of a combination of a new di- agnostic code for intracranial hemorrhage and a new procedure code for brain imaging, new intracranial hemorrhages by recurrent head injury could be misidentified as delayed hemorrhages. Therefore, it is possible that the risk factors identified in this study are simply risk factors for re- current falls and subsequent new intracranial hemorrhage. Second, al- though the sample size was large, the number of users on antithrombotic agents were small, especially for those who were taking clopidogrel and vitamin-K antagonists. Third, further adjusts for the mechanisms of trauma and the severity of the injury were not consid- ered. Lastly, the prediction models presented in this study need to be ex- ternally validated in prospectively collected cohorts, preferentially in multi-center study setting. Despite these limitations, this population- based study has several strengths. Because of its large sample size, it allowed us to identify new risk factors for the delayed intracranial hem- orrhage in real-world ED setting. The more diverse study population not restricted to those with anticoagulant users can also be useful in obtaining relatively unbiased estimates of the risk of delayed intracrani- al hemorrhage. Finally, a risk stratification system for the prediction of delayed intracranial hemorrhage could be constructed which has not been available in the previous studies [6,7].

In conclusion, we found old age, associated injuries including cranio- facial fracture and any neck injury and comorbidities such as diabetes mellitus and hypertension are the independent risk factors of delayed intracranial hemorrhage and risk stratification based on them showed reasonable performance.

Source of support

This study was supported by Seoul National University Bundang Hospital (SNUBH) grant 02-2014-045.

Conflicts of interest

No authors have any conflicts of interest.

Acknowledgments

The authors thank the Division of Statistics in Medical Research Col- laborating Center at Seoul National University Bundang Hospital for sta- tistical analyses.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ajem.2017.08.009.

References

  1. Fu TS, Jing R, Fu WW, Cusimano MD. Epidemiological trends of traumatic brain inju- ry identified in the Emergency Department in a Publicly-Insured Population, 2002-2010. PLoS One 2016;11:e0145469.
  2. Whiteneck GG, Cuthbert JP, Corrigan JD, Bogner JA. Prevalence of self-reported life- time history of traumatic brain injury and associated disability: a statewide popula- tion-based survey. J Head Trauma Rehabil 2016;31:E55-2.
  3. Hamilton M, Mrazik M, Johnson DW. Incidence of delayed intracranial hemorrhage in children after uncomplicated minor head injuries. Pediatrics 2010;126:e33-9.
  4. Matsuda W, Sugimoto K, Sato N, Watanabe T, Fujimoto A, Matsumura A. Delayed onset of posttraumatic acute subdural hematoma after mild head injury with nor- mal computed tomography: a case report and brief review. J Trauma 2008;65: 461-3.
  5. Vos PE, Battistin L, Birbamer G, et al. EFNS guideline on mild traumatic brain injury: report of an EFNS task force. Eur J Neurol 2002;9:207-19.
  6. Miller J, Lieberman L, Nahab B, et al. Delayed intracranial hemorrhage in the anticoagulated patient: a systematic review. J Trauma Acute Care Surg 2015;79: 310-3.
  7. Shabani S, Nguyen HS, Doan N, Baisden JL. Case report and review of literature of de- layed acute subdural hematoma. World Neurosurg 2016;96:66-71.
  8. Schoonman GG, Bakker DP, Jellema K. Low risk of late Intracranial complications in mild traumatic brain injury patients using Oral anticoagulation after an initial nor- mal brain computed tomography scan: education instead of hospitalization. Eur J Neurol 2014;21:1021-5.
  9. Li J. Admit all anticoagulated head-injured patients? A million dollars versus your dime. You make the call. Ann Emerg Med 2012;59:457-9.
  10. Swap C, Sidell M, Ogaz R, Sharp A. Risk of delayed Intracerebral hemorrhage in anticoagulated patients after minor head trauma: the role of repeat cranial comput- ed tomography. Perm J 2016;20:14-6.
  11. Folkerson LE, Sloan D, Cotton BA, Holcomb JB, Tomasek JS, Wade CE. Predicting pro- gressive hemorrhagic injury from isolated traumatic brain injury and coagulation. Surgery 2015;158:655-61.
  12. Lee J, Lee JS, Park SH, Shin SA, Kim K. Cohort profile: The National Health insurance service-national sample cohort (NHIS-NSC), South Korea. Int J Epidemiol 2016:dyv319.
  13. The Comprehensive R. Archive network. https://cran.r-project.org/web/packages/ smbinning/smbinning.pdf, Accessed date: 29 May 2017. [Available at, Archied with Webcite].
  14. Kaen A, Jimenez-Roldan L, Arrese I, et al. The value of sequential computed tomog- raphy scanning in anticoagulated patients suffering from Minor head injury. J Trau- ma 2010;68:895-8.
  15. Peck KA, Sise CB, Shackford SR, et al. Delayed intracranial hemorrhage after blunt trauma: are patients on preinjury anticoagulants and prescription antiplatelet agents at risk? J Trauma 2011;71:1600-4.
  16. Nishijima DK, Offerman SR, Ballard DW, et al. Immediate and delayed traumatic in- tracranial hemorrhage in patients with head trauma and preinjury warfarin or clopidogrel use. Ann Emerg Med 2012;59 [460-468 e461-467].
  17. Menditto VG, Lucci M, Polonara S, Pomponio G, Gabrielli A. Management of minor head injury in patients receiving Oral anticoagulant therapy: a prospective study of a 24-hour observation protocol. Ann Emerg Med 2012;59:451-5.
  18. McCammack KC, Sadler C, Guo Y, Ramaswamy RS, Farid N. Routine repeat head CT may not be indicated in patients on anticoagulant/Antiplatelet therapy following mild traumatic brain injury. West J Emerg Med 2015;16:43-9.
  19. Chauny JM, Marquis M, Bernard F, et al. Risk of delayed intracranial hemorrhage in anticoagulated patients with mild traumatic brain injury: systematic review and meta-analysis. J Emerg Med 2016;51:519-28.
  20. Keenan HT, BrundAge SI, Thompson DC, Maier RV, Rivara FP. Does the face protect the brain? A case-control study of traumatic brain injury and facial fractures. Arch Surg 1999;134:14-7 [Chicago, Ill.: 1960].
  21. Hemphill III JC, Greenberg SM, Anderson CS, et al. Guidelines for the management of spontaneous intracerebral hemorrhage: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2015;46:2032-60.
  22. Boulanger M, Poon MT, Wild SH, Al-Shahi Salman R. Association between diabetes mellitus and the occurrence and outcome of intracerebral hemorrhage. Neurology 2016;87:870-8.