Article, Surgery

Opportunity to reduce transfer of patients with mild traumatic brain injury and intracranial hemorrhage to a Level 1 trauma center

a b s t r a c t

Objective: Current guidelines do not address the disposition of patients with mild Traumatic brain injury and resultant Intracranial Hemorrhage . Emergency medicine clinicians working in hospitals without neurosur- gery coverage typically transfer patients with both to a trauma center with neurosurgery capability. Evidence is accruing which demonstrates that the risk of neurologic decompensation depends on the type of ICH and as a re- sult, not every patient may need to be transferred. The purpose of this study was to identify risk factors for admis- sion among patients with mild TBI and ICH who were transferred from a community hospital to the emergency department (ED) of a Level 1 trauma center.

Methods: Study subjects were patients >=18 years of age who were transferred from a community hospital to the ED of an urban, academic Level 1 trauma center between April 1, 2015 and March 31, 2016, and with an isolated traumatic ICH. Patients who had an epidural hematoma, were deemed to require a trauma center’s level of ser- vice, were found to have non-traumatic ICHs, or had a Glasgow Coma Scale of b 13 were excluded. Using a mul- tivariable logistic regression model, we sought to determine patient factors and computed tomography findings which were associated with admission (to the floor, intensive care unit, or operating room with neuro- surgery) of the Level 1 trauma center.

Results: 644 Transferred patients were identified; 205 remained eligible after exclusion criteria. Presence of war- farin (odds ratio [OR] 4.09, 95% Confidence Interval [CI] 1.64, 10.25, p = 0.0026) and a subdural hematoma

>=1 cm (OR 6.28, 95% CI 1.24, 31.71, p = 0.0263) were independently statistically significant factors predicting ad- mission. Age, sex, GCS, presence of neurologic deficit, aspirin use, clopidogrel use, SDH b 1 cm, IPH, and SAH were each independently not significant predictive factors of an admission.

Conclusions: After controlling for factors, transferred patients with mild TBI with a SDH >=1 cm or on warfarin have a higher odds ratio of requiring inpatient admission to a Level 1 trauma center. While these patients may require admission, there may be opportunities to develop and study a low risk traumatic intracranial hemorrhage proto- col, which keeps a subgroup of patients with a mild TBI and resultant ICH at community hospitals with access to a nearby Level 1 trauma center.

(C) 2017

  1. Introduction

Annually, an estimated 1.5 million people in the United States (U.S.) sustain a non-fatal head injury, 80% of which are considered mild

* Corresponding author at: Department of Emergency Medicine, Massachusetts General Hospital, 0 Emerson Place, Suite 3B, Boston, MA 02114, United States.

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

Traumatic brain injuries (TBI) [1,2]. About 10% of those with TBI are found to have an Intracranial hemorrhage [3]. While American College of Emergency Physicians guidelines exist for managing patients with mild TBI without ICH, no current U.S. guidelines address what to do with patients who have mild TBI, defined as Glasgow Coma Scale (GCS) 13-15 [1], with ICH [4]. As a result, there is wide variation in care of pa- tients who suffered a mild TBI with CT evidence of ICH [5,6], and clini- cians have little data to assist in determining the most appropriate disposition.

0735-6757/(C) 2017

1282 B.J. Yun et al. / American Journal of Emergency Medicine 35 (2017) 12811284

Due to the potential for poor outcomes in this patient population [3], emergency department (ED) clinicians working in hospitals without neurosurgery access typically transfer such patients to a hospital with neurosurgery availability [7]. However, mounting evidence shows that intracranial hemorrhage in mild TBI is not a homogeneous disease. Compared to patients with mild TBI and subdural hemorrhages, pa- tients with mild TBI with isolated Subarachnoid hemorrhages and cere- bral contusions are less likely to need neurosurgical intervention [6,8,9]. As a result, it is unlikely that every patient with mild TBI and evidence of ICH requires transfer to a tertiary care center.

Therefore, the purpose of this study was to identify factors predicting admission (to the floor, the intensive care unit, or operating room with neurosurgery) among patients with mild TBI and ICH who were transferred from a community hospital to the ED of a Level 1 trau- ma center.

  1. Methods
    1. Study design

This was a Health Insurance Portability and Accountability Act-com- pliant, retrospective cohort study, approved by the study site’s Institu- tional Review Board.

Study setting and population

The study was performed at an urban, 999-bed quaternary care aca- demic center with an ED hosting an ACGME-accredited emergency medicine residency, carrying both a Level 1 Adult and Pediatric trauma center designation, and conducting approximately 109,000 annual visits. The ED also has observation unit (EDOU) capability, composed of 32 beds staffed by both nurse practitioners and physician assistants with emergency medicine physician oversight. At our institution, all pa- tients with a traumatic ICH receive a neuroSurgical consultation. For pa- tients with non-operative traumatic ICH, emergency medicine physicians decide the disposition in consultation with neurosurgeons, which may include discharge, placement in EDOU, admission to the floor, or admission to the intensive care unit (ICU). For patients with an isolated non-operative ICH who require an admission to the floor, the neurosurgery, trauma surgery, and neurology services admit pa- tients on a scheduled rotation.

Using our ED reporting system, we searched for patients >= 18 years old who were transferred to our ED from surrounding community hos- pitals between April 1, 2015 and March 31, 2016 with a chief complaint of traumatic ICH, ICH, subdural hematoma (SDH), or subarachnoid hem- orrhage (SAH). We sought to include patients with an isolated traumat- ic ICH. Patients who had a traumatic ICH with a minor orthopedic injury not requiring surgery were included. We excluded epidural hematoma

Outcome measure

Our composite endpoint was admission to the floor, ICU, or operat- ing room (OR) for a neurosurgical operation from the ED.

Data analysis

Baseline characteristics were summarized using descriptive statisti- cal methods. We identified factors potentially predicting admission (floor, ICU or OR), our primary outcome measure, based on our prior knowledge and prior mild TBI studies [6,8], and used them to create a logistic regression model. We identified the following factors: age, sex, initial GCS, presence of neurologic deficit, warfarin use, clopidogrel use, aspirin use, and presence of SDH (b 1 cm or >= 1 cm), SAH, or IPH. We chose an alpha of 0.05 as the cutoff for significance. We used SAS

9.4 (SAS Institute Inc., Cary, North Carolina) for the statistical analysis [11].

  1. Results
    1. Study population

A total of 644 patients were identified during the study period; 2 were excluded as they had an epidural hematoma and 174 were exclud- ed as they suffered multiple traumatic injuries requiring additional sur- gical subspecialty services, had a GCS b 13, or arrived intubated from the community hospital. A further 263 were excluded as they were found to have non-traumatic ICHs. Baseline characteristics of the 205 patients who remained eligible and comprise the study population are presented in Table 1.

Patient disposition

98 (48%) patients ultimately required admission, including 12 who were admitted to the floor after a stay in the EDOU. All 8 of the patients who went to the operating room had a SDH >= 1 cm. Additional Patient dispositions are presented in Fig. 1.

Predictors for admission

In multivariable regression modeling, warfarin use (odds ratio [OR] 4.09, 95% confidence interval [CI] 1.64, 10.25, p = 0.0026) and a SDH

>= 1 cm (OR 6.28, 95% CI 1.24, 31.71, p = 0.0263) were independently

Table 1

Baseline characteristics.

as it is associated with high mortality [10]. Additional exclusion criteria

were patients deemed to require a trauma center’s level of service (pa- tient required additional surgical subspecialty services, such as oral maxillofacial surgery, plastic surgery, ophthalmology, or operative or- thopedic services), were intubated at the community hospital, were found to have non-traumatic ICHs, or had a GCS of b 13.

2.3. Data collection

For each patient, we reviewed their electronic medical record and collected age, sex, initial ED GCS, presence of neurologic deficit (motor or sensory), anticoagulation use, anti-platelet therapy, Computed To- mography (CT) head findings, and disposition from the ED. We record- ed CT findings based on the final neuroradiologist report. Imaging findings included SDH (b 1 cm or >= 1 cm), SAH, or intraparenchymal hemorrhage (IPH). We also recorded the subsequent disposition after placement in the EDOU (discharge, floor, ICU, OR).

Variable Value (n = 205)

Mean age (standard deviation) 71 years [17]

Male (%) 95 (46%)

GCS (%)

13 2 (1%)

14 13 (7%)

15 190 (92%)

Neuro deficit 18 (9%)

ASA use 77 (38%)

Clopidogrel use 9 (4%)

Warfarin use 33 (16%)

Novel Oral anticoagulation use 0 (0%) CT findingsa


b1 cm 97

>= 1 cm 23

Subarachnoid 83

Intraparenchymal 18

a As patients can have more than one CT finding, total is N 205.

B.J. Yun et al. / American Journal of Emergency Medicine 35 (2017) 12811284 1283

Fig. 1. Cohort flow diagram of disposition. EDOU = Emergency department observation unit, ICU = Intensive care unit, OR = Operating room.

statistically significant predictive factors for admission. Age, sex, GCS, presence of neurologic deficit, aspirin use, clopidogrel use, SDH b 1 cm, IPH, and SAH were each independently not significant predictive factors of an admission (Table 2).

  1. Discussion

Our study’s goal was to define which patients with a mild TBI and re- sultant ICH who were transferred from a community hospital may ap- propriately require a Level 1 trauma center’s ability to provide a higher level of resources, which includes intervention by neurosurgery, admission to the floor, or admission to the intensive care unit. In our Level 1 trauma center, warfarin use and SDH >=1 cm were each indepen- dently associated with a higher odds ratio of admission. Our work con- tributes to the understanding of the resources potentially needed to care for patients with mild TBI and intracranial hemorrhage [3,5,7,9, 12-19].

Table 2

Odds ratio of admission from logistic regression.

Parameter Odds ratio 95% confidence interval p-Value

First, as noted in the methods, after controlling for age, sex, GCS, presence of neurologic deficit, warfarin use, aspirin use, clopidogrel use, SDH b 1 cm, IPH, and SAH, the odds of admission for patients with a SDH >= 1 cm were 528% higher than in patients without a SDH >= 1 cm. In addition, every patient in our study that went to the operating room with neurosurgery from the ED had SDH >= 1 cm. Prior studies have shown that larger SDH are associated with worse outcomes. For example, while it did not control for factors, Albertine et al. in a retro- spective study found that patients with a SDH volume of >= 10cm3 were associated with neurological decline and worse clinical outcomes [20]. Additionally, in an effort to determine who would require an ICU admis- sion, Washington et al. found that ICH volume N 10 ml was associated with a higher risk for hemorrhagic progression on repeat CT [14].

Our study also builds on the work of previous mild TBI research demonstrating that clinical outcomes vary depending on the type of ICH. Borczuk et al. found that patients with a GCS of 15 and traumatic subarachnoid hemorrhage had lower risk for clinical deterioration, neu- rosurgical intervention, or worsening CT results [8]. Using the National Trauma Data Bank, Sweeney et al. found that patients with subdural hemorrhages and epidural hemorrhages compared to patients with in- tracranial contusions were more likely to require neurosurgical inter- vention [6].

Second, warfarin use was independently associated with a higher



0.99, 1.03


odds ratio of admission. The odds of admission for patients on warfarin



0.64, 2.27


were 309% higher than in patients not on warfarin. Prior studies also

GCS 0.33 0.10, 1.13 0.0770

Presence of neurologic deficit 1.63 0.49, 5.42 0.4287

Aspirin use 0.66 0.33, 1.31 0.2292

Clopidogrel use 0.57 0.11, 2.92 0.4980

Warfarin use 4.09 1.64, 10.25 0.0026? Subdural hemorrhage


b1 cm




0.27, 2.07



>= 1 cm


1.24, 31.71


showed that patients with a traumatic ICH on anticoagulation are at a higher risk for poor clinical outcomes [21]. Schaller et al. found that adult patients not on anticoagulant therapy may not require a period of hospitalization or observation [12]. Patients on warfarin may require reversal of an elevated International normalized ratio and moni- toring for signs of hemorrhagic progression [14]. In addition after stabi- lization, patients require counseling regarding the risks and benefits of

Subarachnoid hemorrhage


0.33, 2.31


restarting anticoagulation in the future and, if resuming anticoagulation,

Intraparenchymal hemorrhage


0.53, 6.31


the optimal initiation timing. As a result, these patients may require a

* Statistically significant.

period of Close monitoring at a trauma center.

1284 B.J. Yun et al. / American Journal of Emergency Medicine 35 (2017) 12811284

Finally, these findings highlight a potential opportunity to reduce unnecessary transfers with the implementation of telemedicine ser- vices. A mild TBI with ICH protocol targeting patients who are low risk for deterioration could be created to keep appropriate patients at the community hospital. Patients with SDH >= 1 cm or on warfarin should al- ways be transferred. This strategy could potentially leverage access to higher acuity care, while optimizing appropriate site of care for lower risk patients. For example, Ashkenazi et al. found that selected patients with traumatic intracranial hemorrhage could be safely managed at the local hospital with neurosurgical consultation via telemedicine [15]. Levy et al. found that at their Level III trauma hospital without neurosur- gical capabilities, they were able to safely implement a nontransfer pro- tocol for mild TBI with either a small SAH, punctuate cerebral contusion, or very small SDH without mass effect [22]. CT scans and clinical status were reviewed by an on-call neurosurgeon from the Level 1 trauma cen- ter [22]. Our study supports these findings, but a larger prospective study will need to validate the results.

  1. Limitations

This study has several limitations in addition to its retrospective de- sign. First, we did not examine the post-discharge outcomes of the pa- tients. However, given this was a descriptive study of resource use, this analysis was outside the scope of our research question. Second, be- cause this study involved a single Level 1 trauma center in which all pa- tients transferred for ICH receive neurosurgical consultation, the ability to generalize our results may be limited to other hospitals with similar characteristics. In addition as our institution is one of the larger Level 1 trauma centers, it is unlikely that the decision to transfer to our insti- tution verses other Level 1 trauma centers in the area was correlated with patients’ severity. Finally another limitation was that our cohort did not include any subjects on novel anticoagulants.

  1. Conclusions

After controlling for selected factors, transferred patients with mild TBI with a SDH >= 1 cm or on warfarin have a higher odds ratio of requir- ing an admission at a Level 1 trauma center. While patients with a SDH

>= 1 cm or on warfarin require admission, there may be opportunities to develop a protocol to manage a subgroup of patients with a mild TBI and resultant ICH at community hospitals with access to a nearby Level 1 Trauma Center. Future research should evaluate a low risk traumatic in- tracranial hemorrhage transfer protocol among community hospitals and a Level 1 trauma center.

Source of support

This research did not receive any specific grant from funding agen- cies in the public, commercials, or not-for-profit sectors.




  1. Carroll LJ, Cassidy JD, Holm L, Kraus J, Coronado VG. Methodological issues and re- search recommendations for mild traumatic brain injury: the WHO Collaborating

Centre Task Force on mild Traumatic Brain Injury. J Rehabil Med 2004;36(Suppl. 43):113-25.

  1. Sosin DM, Sniezek JE, Thurman DJ. Incidence of mild and moderate brain injury in the United States, 1991. Brain Inj 1996;10(1):47-54. 026990596124719.
  2. Kreitzer N, Lyons MS, Hart K, et al. Repeat neuroimaging of mild traumatic brain-in- jured patients with acute traumatic intracranial hemorrhage: clinical outcomes and radiographic features. Acad Emerg Med 2014;21(10):1083-91. 1111/acem.12479.
  3. Jagoda AS, Bazarian JJ, Bruns JJ, et al. Clinical policy: neuroimaging and decisionmaking in adult mild traumatic brain injury in the acute setting. Ann Emerg Med 2008;52(6):714-48. 08.021.
  4. Nishijima DK, Haukoos JS, Newgard CD, Staudenmayer K, White N, Slattery D, et al. Variability of ICU use in adult patients with minor traumatic intracranial hemorrhage. Ann Emerg Med 2013;61(5):509-17 (e4) 10.1016/j.annemergmed. 2012.08.024.
  5. Sweeney TE, Salles A, Harris O a, Spain D a, Staudenmayer KL. Prediction of neuro- surgical intervention after mild traumatic brain injury using the national trauma data bank. World J Emerg Surg 2015;10(1):23. 015-0017-6.
  6. Bee TK, Magnotti LJ, Croce MA, Maish GO, Minard G, Schroeppel TJ, et al. Necessity of repeat head CT and ICU monitoring in patients with minimal brain injury. J Trauma 2009;66(4):1015-8.
  7. Borczuk P, Penn J, Peak D, Chang Y. Patients with traumatic subarachnoid hemor- rhage are at low risk for deterioration or neurosurgical intervention. J Trauma Acute Care Surg 2013;74(6):1504-9. 0b013e31829215cf.
  8. Phelan HA, Richter AA, Scott WW, et al. Does isolated traumatic subarachnoid hem- orrhage merit a lower intensity level of observation than other traumatic brain inju- ry? J Neurotrauma 2014;31(20):1733-6.
  9. Bir SC, Maiti TK, Ambekar S, Nanda A. Incidence, hospital costs and in-hospital mor- tality rates of epidural hematoma in the United States. Clin Neurol Neurosurg 2015; 138:99-103.
  10. SAS Institute Inc. SAS for Windows. Release 9.4; 2012(Cary, NC).
  11. Schaller B, Evangelopoulos DS, Muller C, Martinolli L, Pouljadoff MP, Zimmermann H, et al. Do we really need 24-h observation for patients with minimal brain injury and small Intracranial bleeding? The Bernese Trauma Unit Protocol. Emerg Med J 2010;27(7):537-9.
  12. 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(6):451-5. http://dx.doi. org/10.1016/j.annemergmed.2011.12.003.
  13. Washington CW, Grubb RL. Are routine repeat imaging and intensive care unit ad- mission necessary in mild traumatic brain injury? J Neurosurg 2012;116(3): 549-57.
  14. Ashkenazi I, Haspel J, Alfici R, Kessel B, Khashan T, Oren M. Effect of teleradiology upon pattern of transfer of head injured patients from a rural general hospital to a neurosurgical referral centre. Emerg Med J 2007;24(8):550-2. 1136/emj.2006.044461.
  15. AbdelFattah KR, Eastman AL, Aldy KN, Wolf SE, Minei JP, Scott WW, et al. A prospec- tive evaluation of the use of routine repeat cranial CT scans in patients with intracra- nial hemorrhage and GCS score of 13 to 15. J Trauma Acute Care Surg 2012;73(3): 685-8.
  16. Nayak NV, Medina B, Patel K, Homnick AT, Mohr AM, Livingston DH, et al. Neurologic outcome of minimal head injury patients managed with or without a routine repeat head computed tomography. J Trauma Acute Care Surg 2013;75(2):273-8. http://
  17. Velmahos GC, Gervasini A, Petrovick L, Dorer DJ, Doran ME, Spaniolas K, et al. Rou- tine repeat head CT for minimal head injury is unnecessary. J Trauma 2006;60(3): 494 -499-501 10.1097/01.ta.0000203546.14824.0d.
  18. Fabbri a, Servadei F, Marchesini G, Stein SC, Vandelli a. Observational approach to subjects with mild-to-moderate head injury and initial non-neurosurgical lesions. J Neurol Neurosurg Psychiatry 2008;79(10):1180-5. 2007.135178.
  19. Albertine P, Borofsky S, Brown D, Patel S, Lee W, Caputy A, et al. Small subdural hemorrhages: is routine intensive care unit admission necessary? Am J Emerg Med 2016;34(3):521-4.
  20. Cohen DB, Rinker C, Wilberger JE. Traumatic brain injury in Anticoagulated patients. J

Trauma 2006;60(3):553-7.

  1. Levy AS, Orlando A, Salottolo K, Mains CW, Bar-Or D. Outcomes of a nontransfer pro- tocol for mild traumatic brain injury with abnormal head computed tomography in a Rural hospital setting. World Neurosurg 2014;82(1-2):e319-23. 10.1016/j.wneu.2013.11.008.