Risk Factors for Post-Traumatic Massive Cerebral Infarction Secondary to Space-Occupying Epidural Hematoma

Abstract

Post-traumatic massive cerebral infarction (MCI) is a fatal complication of concurrent epidural hematoma (EDH) and brain herniation that commonly requires an aggressive decompressive craniectomy. The risk factors and surgical indications of MCI have not been fully elucidated. In this retrospective study, post-traumatic MCI was diagnosed in 32 of 176 patients. The performance of a decompressive craniectomy simultaneously with the initial hematoma-evacuation surgery improved their functional outcomes, compared with delayed surgery (on the 6-month Extended Glasgow Outcome Scale, 5.6±1.5 vs. 3.4±0.6; p<0.001). Significantly increased risks for MCI were observed in patients with an EDH at a transtemporal location (adjusted odds ratio [OR], 16.48; p=0.003), an EDH larger than 100 mL in volume (OR, 7.04; p=0.001), preoperative shock for longer than 30 min (OR, 13.78; p=0.002), bilateral mydriasis (OR, 7.08; p=0.004), preoperative brain herniation for longer than 90 min (OR, 6.41; p<0.001), and a Glasgow Coma Score of 3–5 points (OR, 2.86; p<0.053). Multi-variate logistic regression analysis revealed no significant association between post-traumatic MCI and age, gender, mid-line shift, Rotterdam computed tomography score, intraoperative hypotension, or serum concentrations of sodium or glucose. Incidence of post-traumatic MCI increased from 16.4% in those having any two of the six risk factors to 47.7% in those having any three or more of the six risk factors (p<0.001). Patients with concurrent EDH and brain herniation exhibited an increased risk for post-traumatic MCI with the accumulation of several critical clinical factors. Early decompressive craniectomy based on accurate risk estimation is recommended in efforts to improve patient functional outcomes.

Introduction

Acute epidural hematoma (EDH) because of traumatic brain injury accounts for approximately 21.5–50% of intracranial hematomas and is characterized by rapid disease progression when a brain herniation is found to exist concurrently, leading to relatively high morbidity. If the EDH-produced mass effect is not relieved in time, the overall outcome is poor. Ordinary therapeutic strategies include emergent twist-drill drainage and craniotomy surgery to evacuate the hematoma; these strategies principally aim to decrease intracranial pressure (ICP) and promote the repositioning of the brain herniation.^1,2 There is a set of patients who experience clinical deterioration after an initial hematoma-evacuation craniotomy because of secondary brain injuries, including massive cerebral infarction (MCI). Additional decompressive craniectomy is recommended as soon as possible.^3,4 Decompressive craniectomy allows for the extracranial expansion of the infarcted and oedematous brain parenchyma. This procedure alleviates the physical oppression of the ventricular system to maintain a normal circulation of cerebrospinal fluid, the blood circulation system to ensure an essential blood supply and venous drainage, and the brainstem structures to avoid the interruption of basic life support functions.^5,6 Because the incidence of MCI secondary to acute EDH is relatively low, which has been well recognized in clinical practice with solid epidemiological evidence, the removal of the bone flap is not necessary in the majority of patients with concurrent EDH and brain herniation. Inappropriate decompressive craniectomy performed simultaneously with initial hematoma-evacuation surgery brings unavoidable complications, such as abnormal hemodynamics and subsequent cerebral necrosis and infarction as well as a need for cranioplasty. Therefore, there is a clinical rationale to investigating the underlying mechanisms and predictive factors of post-traumatic MCI secondary to acute EDH, thereby providing a reliable reference for the decision to remove the bone flap early during initial hematoma-evacuation surgery.

Methods

Inclusion and exclusion criteria

This retrospective study was approved by the local ethics committee in our hospital and enrolled a set of 176 patients with supratentorial brain herniation caused by traumatic acute EDH. Clinical data were collected from January 2005 to January 2012 in our neurosurgery department. Inclusion criteria included the following: 1) computed tomography (CT) scan-evidenced traumatic EDH and concurrent herniation and 2) craniotomy or craniectomy for hematoma evacuation performed on patients in our department. Exclusion criteria included the following: 1) In addition to EDH, the patients had coexistent severe cerebral contusion and laceration or subdural hematoma (SDH) that had a marked mass effect, synergistically contributing to brain herniation; 2) patients who had injury of the oculomotor nerve resulting from a basal skull fracture; 3) patients who had diabetes mellitus or other diseases that caused a hypercoagulative state; 4) patients who were older than 70 years; 5) patients who had combined severe heart and lung disease; and 6) patients who had incomplete clinical data, especially the basic records of the first-aid neurological examinations.

Radiographic evaluation

Rotterdam CT scores of patients were calculated on the basis of the most recent preoperative CT image, as described previously. EDH volume and post-traumatic MCI were determined with a calculation (length×width×height)/2 in the largest area of the EDH/MCI according to Coniglobus' formula. The mid-line shift (MLS) was determined on CT images (on the septum pellucidum plane) by measuring the perpendicular dimension from the center of the septum pellucidum to the virtual line connecting the center inner ridges of the frontal and occipital skull bone.

The primary endpoint of this study was traumatic MCI, which was primarily diagnosed by the CT examination according to the following criteria⁷: 1) No obvious cerebral infarction was detected in the initial CT examination upon admission; 2) after initial hematoma-evacuation surgery, the CT re-examination depicted a massive low-dense area that occupied more than two thirds of the space or involved more than 2 lobes; and 3) the maximum diameter of the cerebral infarction was longer than 4 cm. Brain edema was differentiated from MCI predominantly by the location and measurement of the CT value. Brain edema mainly occurred in the areas surrounding the cerebral contusions and had a relatively higher CT value, compared with MCI. The CT examination was performed ordinarily before surgery and 1, 3, and 7 days after surgery. Diffusion-weighted imaging/apparent diffusion coefficient/magnetic resonance imaging (DWI/ADC-MRI) was applied to confirm the MCI diagnosis. These radiographic images were interpreted by two skilled double-blinded neuroradiologists.

Treatments and surgical decisions

After admission, all patients received conventional therapies such as management of ICP and blood pressure and were given emergency surgical interventions according to their disease conditions and the decision of their lineal relatives. Informed consent was obtained before the surgical procedures. According to the clinical condition of the individual patients, three skilled surgeons performed the emergent hematoma-evacuation craniotomy in all 176 of the patients. Among them, 13 patients received synchronous decompressive craniectomy by the removal of the bone flaps, whose size was designed approximately according to, and in approximation of, the maximal margin of the EDH. They received individualized hyperbaric oxygen treatment and dehydration therapy and did not undergo secondary craniectomy to enlarge the skull defect. In these 13 patients, post-traumatic MCI was highly suspected intraoperatively and was rationalized on the basis of the following: 1) Preoperative CT examinations indicated transtemporal EDH with large volume and a great mass effect, severe brain herniation depicted by the absence of the basal cistern, and a more than 10-mm MLS; 2) intraoperative observation of tense dura despite evacuation of the EDH, attenuated cerebral pulsation, and pale cortical surface and/or tense cerebral parenchyma without delayed contralateral or distant-part–located SDH/EDH revealed by an intraoperative CT scan; and, more important, 3) pre- and/or intraoperative CT examinations indicated occurrence of multiple low-dense ischemic lesions in the temporal lobe, basal ganglia, and occipital lobe. The above-mentioned craniectomy was performed with decompressive duraplasty.

Patient prognosis

The secondary endpoint was indicated by the long-term functional outcomes, including overall mortality and the score on the Extended Glasgow Outcome Scale (GOS-E),⁸ which was scored as follows: 1) death; 2) persistent vegetative state; 3) lower severe disability; 4) upper severe disability (stratum 3 and 4 were considered as severe disability, with permanent requirement for help with daily living); 5) lower moderate disability; 6) upper moderate disability (stratum 5 and 6 were considered as mild disability, without a need for assistance in everyday life, that might, however, require special equipment for employment); 7) lower good recovery; and 8) upper good recovery (stratum 7 and 8 were considered as good recovery). Assessments of clinical outcomes were performed by professional neurosurgeons who participated in this study at the time of discharge and 1, 3, 6, and 12 months after discharge.

Clinical indices involved in the regression analysis

This study used 13 clinical indices for the investigation of potential risk factors for post-traumatic MCI. These indices included the age, gender, volume, and anatomical location of EDH, duration and extent of preoperative brain herniation, Glasgow Coma Scale (GCS) score, MLS, Rotterdam CT score, concentrations of blood glucose and blood sodium, occurrence of preoperative shock, and intraoperative hypotension. According to their medical characteristics and statistical distributions, these indices were organized into unordered and ordinal categorical variables, respectively.

Statistical analysis

In the preliminary selection by univariate analysis, the baseline characteristics between the groups were compared using Pearson's chi-square test for categorical variables (two dependent 4-fold table samples and constituent ratio samples). Kruskal-Wallis' nonparametric test was performed for two dependent ranked samples to determine the difference between the two study groups. The statistically significant indices (p<0.2) determined by the univariate analyses were then enrolled into the unconditional multi-variate logistic step-wise regression analysis. The “p to enter” cut-off value was specified as 0.10, and the “p to stay” cut-off value was specified as 0.15.

Results

Patient demographics and clinical characteristics

According to the indicated patient selection criteria, a cohort of 176 patients with traumatic acute EDH and complicated supratentorial herniation were enrolled in this study. Injury mechanisms included traffic accident (134 cases), blunt brain injury (19 cases), and fall injury (23 cases). Of the 176 patients enrolled in the study, there were 124 males and 52 females. Mean age was 38.8 years (range, 2–70), and mean time from occurrence of the brain herniation to the initial hematoma-evacuation surgery was 72.2±34.8 min (range, 24–186). The EDH was located in the temporal region in 23 cases, the frontal region in 21 cases, the tempro-parietal region in 43 cases, the tempro-occipital region in 37 cases, and other areas in 52 cases. Mean GCS score upon admission was 7.2 (range, 3–13), and mean volume of the EDH was 81.5 mL (range, 40–158), which was less than 60 mL in 42 cases, 60–100 mL in 98 cases, and greater than 100 mL in 36 cases. The mean MLS was 12.4 mm (range, 4–20). Unilateral mydriasis was observed in 135 cases, whereas bilateral mydriasis was observed in 41 cases. Preoperative shock with systolic pressure of less than 90 mm Hg for longer than 30 min was recorded in 23 cases. Intraoperative hypotension with systolic pressure of less than 90 mm Hg for longer than 30 min was determined in 14 cases. Mean concentration of blood sodium was 139.6 mmol/L (range, 124.2–160.2). Hypernatremia was determined in 42 cases, whereas hyponatremia was determined in 43 cases. Mean concentration of blood glucose was 9.5 mmol/L (range, 4.5–23.5). Hyperglycemia was determined in 135 cases, among which 47 cases had a >11.1-mmol/L blood glucose and 41 cases had a <7.0-mmol/L blood glucose.

Incidence and clinical characteristics of traumatic massive cerebral infarction

In the study population, 32 patients were diagnosed with post-traumatic MCI and thus received additional decompressive craniectomy and duraplasty. Incidence of post-traumatic MCI in the indicated patients was approximately 18.2% (32 of 176; 95% confidence interval [CI], 12.5–23.9). Clinical characteristics, including volume and distribution of the vascular territory of infarctions are summarized in Table 1.

Table 1.

Clinical Characteristics of Post-Traumatic MCI ^a

Clinical parameters	Left hemispheric (%)	Right hemispheric (%)	Bilateral hemispheric (%)
Volume, mL
<100	8 (25.0)	8 (25.0)	0 (0)
100–150	4 (12.5)	3 (9.4)	0 (0)
>150	4 (12.5)	2 (6.3)	3 (9.4)
Vascular territory
ACA	1 (3.2)	2 (6.3)	0 (0)
MCA	5 (15.6)	4 (12.5)	0 (0)
PCA	4 (12.5)	2 (6.3)	0 (0)
MCA+ACA	2 (6.3)	1 (3.2)	2 (6.3)
MCA+PCA	3 (9.4)	2 (6.3)	0 (0)
MCA+ACA+PCA	1 (3.2)	2 (6.3)	1 (3.2)

n=32.

MCI, massive cerebral infarction; ACA, anterior cerebral artery; MCA, middle cerebral artery; PCA, posterior cerebral artery.

Radiological evaluation of traumatic massive cerebral infarction

Representative CT images of post-traumatic MCIs are shown in Figure 1. These radiological images are characterized by blocks of low-density areas in the ipsilateral hemicerebrum (31 of 32; 96.9%) that, in 1 case, occurred simultaneously in the contralateral hemicerebrum (1 of 32; 3.1%). CT values of massive cerebral infarctions ranged from 8 to 24 HU, with a mean value of 16 HU. Brain edema was commonly observed surrounding the area of massive cerebral infarction, with a mean CT value of 24 HU (range, 20–32). DWI/ADC-MRI was performed in 24 of the 32 patients to confirm the diagnosis of post-traumatic MCI. The anatomical distribution and amount of ultra-early cerebral infarctions indicated by DWI/ADC-MRI were approximately in accord with the hypodensities revealed by CT images. The missed diagnosis by CT examinations of cerebral infarctions commonly located in the contralateral frontal (1 of 24) and parietal lobe (2 of 24), which presented as the impaired circulation of the anterior cerebral artery and its affiliated perforating branches, and in the caudal thalamus (1 of 24), which presented as the impaired circulation of the central branch of the posterior cerebral artery, resulted from a mechanical compression-induced gross shift. DWI/ADC-MRI depicted many more tiny infarctions near the major area of the cerebral infarctions (6 of 24), compared to the CT examination.

FIG. 1.

Representative computed tomography (CT) images of 3 patients who developed post-traumatic massive cerebral infarction (MCI) and received decompressive craniectomy and duraplasty in combination with hematoma-evacuation craniotomy (case 1) or craniotomy alone (cases 2 and 3) at the time of the initial surgery. The 3 patients presented with a space-occupying epidural hematoma (EDH) with a large volume and obvious mid-line shift (A, H, and O) and had at least three of the six risk factors. Integrated craniotomy and decompressive craniectomy could not absolutely abolish the development of post-traumatic MCI secondary to acute EDH, as evidenced by the CT examinations (B, I, and P) and diffusion-weighted imaging/apparent diffusion coefficient/magnetic resonance imaging (DWI/ADC-MRI) examinations (C, D, J, K, Q, and R). However, it would effectively restrain the aggravation of MCIs (E), compared with delayed decompressive craniectomy until the occurrence of obvious MCIs (L and S). When patients successfully passed through the period of high risk for traumatic brain edema (usually 3–10 days after a head trauma), the pre-existing MCIs would be significantly attenuated in the patients who underwent early decompressive craniectomy (F), compared to those who underwent delayed decompressive craniectomy (M and T). At the 3-month follow-up, the hypodensities on the CT re-examination recovered with no obvious encephalomalacia in the patients who underwent early decompressive craniectomy (G), whereas they still existed with moderate-to-severe encephalomalacia in the patients who underwent delayed decompressive craniectomy, as shown in (N) and (U), respectively.

Univariate analysis

Before the multi-variate analysis for possible risk factors associated with post-traumatic MCI, we preliminarily performed univariate analyses to prevent those meaningless clinical indices from entering the final multi-variate logistic step-wise regression model, thereby strengthening its statistical efficacy (Table 2). It was found that the anatomical distribution and volume of the epidural hematomas, the extent (mydriasis) and duration of the preoperative brain herniation, GCS scores, MLS, Rotterdam CT score, and occurrence of preoperative shock were significantly different (p<0.2) between the patients with and without post-traumatic MCI. These eight clinical indices were included in the subsequent multi-variate analysis. The anatomical distribution of the epidural hematoma varied greatly between the two study arms. Significant differences were found independent of whether the anatomical distribution was stratified in the temporal and nontemporal area (29 of 83 vs. 3 of 61; χ²=12.311; p=0.001) or into the frontal area, the frontotemporal area, the temproparietal area, the tempro-occipital area, and other areas (2 of 19 vs. 4 of 19 vs. 15 of 38 vs. 10 of 26 vs. 1 of 42; χ²=14.213; p=0.001). If we focused on the transtemporal EDHs, which were further subgrouped into the frontotemporal area, temporoparietal area, and temporo-occipital area, the anatomical distribution exhibited no differences between the two study arms (4 of 19 vs. 15 of 38 vs. 10 of 26; χ²=1.093; p=0.579). Age distribution, gender distribution, concentrations of blood sodium and blood glucose, and intraoperative hypotension did not differ between the two study arms. Age was not a differential variable, independent of whether it was stratified into children and teenagers (<=20 years), adults (21–65 years), and elderly patients (>=66 years; 5 of 41 vs. 18 of 70 vs. 9 of 33; p=0.190) or into young and elderly patients (<=20 or >=66 years) versus adult patients (21–65 years; 14 of 74 vs. 18 of 70; χ²=0.248; p=0.619).

Table 2.

Potential Clinical Risk Factors Associated with Post-Traumatic MCI in Patients with Concurrent Epidural Hematoma and Brain Herniation

Clinical factors	Substratifications	MCI (n=32) (%)	Non-MCI (n=144) (%)	p value
Age, years	2∼20	5 (15.6)	41 (28.5)	0.190
	21∼65	18 (56.3)	70 (48.6)
	66∼70	9 (28.1)	33 (22.9)
Gender	Male	22 (68.8)	102 (70.8)	0.815
	Female	10 (31.2)	42 (29.2)
Hematoma location	Nontemporal	2 (6.3)	61 (42.4)	0.001
	Temporal	30 (93.8)	83 (57.6)
Duration of preoperative brain hernia, min	<60	5 (15.6)	69 (47.9)	<0.001
	60∼90	12 (37.5)	49 (34.0)
	>90	15 (46.9)	26 (18.1)
GCS points	<=5	8 (25.0)	10 (6.9)	<0.001
	6–8	20 (62.5)	47 (32.6)
	9∼12	4 (12.5)	74 (51.4)
	>12	0 (0.0)	13 (9.0)
Hematoma volume, mL	<60	2 (6.3)	40 (27.8)	<0.001
	60∼100	14 (43.8)	84 (58.3)
	>100	16 (50.0)	20 (13.9)
Mid-line shift, mm	<10	3 (9.4)	38 (26.4)	0.01
	10∼15	17 (53.1)	77 (53.5)
	>15	12 (37.5)	29 (20.1)
Mydriasis	Unilateral	12 (37.5)	123 (85.4)	<0.001
	Bilateral	20 (62.5)	21 (14.6)
Rotterdam CT score	3	2 (6.3)	50 (34.7)	<0.001
	4	10 (31.3)	73 (50.7)
	5	20 (62.5)	21 (14.6)
Blood sodium, mmol/L	<135	8 (25.0)	35 (24.3)	0.819
	135∼145	17 (53.1)	74 (51.4)
	>145	7 (21.9)	35 (24.3)
Blood glucose, mmol/L	<7	7 (21.9)	34 (23.6)	0.565
	7∼11.1	19 (59.4)	69 (47.9)
	>11.1	6 (18.8)	41 (28.5)
Preoperative shock	Yes	8 (25.0)	15 (10.4)	0.027
	No	24 (75.0)	129 (89.6)
Intraoperative hypotension	Yes	4 (12.5)	10 (6.9)	0.293
	No	28 (87.5)	134 (93.1)

MCI, massive cerebral infarction; GCS, Glasgow Coma Scale; CT, computed tomography.

Multi-variate logistic step-wise regression analysis

The final multi-variate logistic step-wise regression model was well established (–2 log likelihood=77.797; Cox and Snell's R square=0.397; Nagelkerke's R square=0.648; p<0.001) and suggested that the anatomical distribution (temporal location, p=0.003) and volume (p=0.001) of the EDHs, the extent (mydriasis, p=0.004) and duration (p<0.001) of brain herniation, GCS scores (p=0.053), and preoperative shock (p=0.002) were the most significant risk factors associated with the development of post-traumatic MCI. The MLS (p=0.501) and Rotterdam CT scores (p=0.325) were precluded at the last step in the model. It was suggested that the temporal location was the most severe risk factor (adjusted odds ratio [OR], 16.48). Another comparable risk factor was the occurrence of preoperative shock, which had an adjusted OR of 13.78. The detailed adjusted OR values of the other risk factors, as well as their significances, are shown in Table 3. Among the above-mentioned risk factors, the duration of preoperative brain herniations had the highest standardized regression coefficient (b^’=0.849), although it had a relative low OR value of 6.408, compared to those of the temporal location and preoperative shock. This result suggests the importance of its contributions to the establishment of a regression model and particular predictive significance to the development of traumatic MCI.

Table 3.

Multi-Variate Step-Wise Logistic Regression Analysis for Risk Factors Associated with Post-Traumatic MCI in Patients with Concurrent Epidural Hematoma and Brain Herniation

Variables	β	β′	Wald	Df	Exp(β) (95% CI)	Significance
Hematoma location	2.802	0.743	9.036	1	16.478 (2.651–102.417)	0.003
Duration of preoperative brain hernia	1.858	0.849	12.557	1	6.408 (2.294–17.903)	0.000
GCS	−1.056	−0.453	3.728	1	0.348 (0.119–1.016)	0.053
Hematoma volume	1.951	0.726	10.125	1	7.039 (2.116–23.417)	0.001
Mydriasis	1.957	0.457	8.220	1	7.080 (1.858–26.988)	0.004
Occurrence of preoperative shock	2.623	0.489	9.411	1	13.782 (2.579–73.664)	0.002

MCI, massive cerebral infarction; GCS, Glasgow Coma Scale.

Prognosis of patients with post-traumatic massive cerebral infarction

There were no cases of death in this study. All the surviving patients received outpatient or telephone follow-up, with a mean follow-up time of 14 months (range, 7–24). At the 6-month follow-up, there were 5 patients with vegetative status, 11 with severe disability, 34 with mild disability, and 126 with good recovery. Patients with post-traumatic MCI (n=32) survived with a mean GOS-E score of 4.4±0.9, which was significantly worse than those without (n=144; GOS-E 7.4±1.2; p<0.001). Among the MCI patients, 13 patients who received a synchronous decompressive craniectomy by the removal of the bone flaps had a GOS-E of 5.6±1.5, whereas 19 who received a delayed decompressive craniectomy until the obvious neurological deterioration and CT scan-confirmed MCI had a GOS-E of 3.4±0.6. This finding suggests the valuable benefits of assessing the neurological function from the early diagnosis onward and of early surgical interventions (p<0.001).

Discussion

Normally, patients with traumatic acute EDH have a good recovery, with a mean length of hospital stay of less than 7 days in 50% of the cases if the hematoma could be evacuated on a timely basis.⁹ Large hematomas might result in brain herniation as early as 3–24 h after head trauma,¹⁰ which induces a gross shift of brain tissues and persistent suppression of the cerebral blood supply and venous drainage and, consequently, leading to the development of post-traumatic MCI and accompanying brain edema requiring intensive medical care or a decompressive craniectomy. Given that repeated surgeries and decompressive craniectomy bring accumulated risks of surgical infections and worsening prognoses in these patients, an integrated surgical strategy combining hematoma-evacuation craniotomy with decompressive craniectomy based on the accurate prediction of post-traumatic MCI has an advantage over ordinary surgery.

The critical pathophysiology of post-traumatic MCI is intracranial hypertension. Standard medical care does not always achieve a satisfactory therapeutic effect, and the overall mortality remains at a high level of 76–80%.¹¹ Decompressive craniectomy and duraplasty permit the outward expansion of infarcted and swollen brain parenchyma and reduce harmful intracranial hypertensions.^6,12 Because aggressive decompressive craniectomy is only needed in cases of post-traumatic MCI, whether the hematoma-evacuation craniotomy should be performed alone or in combination with decompressive craniectomy depends on the accurate evaluation of the high risks of post-traumatic MCI. The univariate analysis and multi-variate logistic regression analysis in this study indicated that the incidence of post-traumatic MCI secondary to acute EDH was approximately 18.2% and was significantly increased from 16.4% in the patients with any two of these six risk factors (transtemproral location, larger than 100 mL in volume, preoperative shock for longer than 30 min, bilateral mydriasis, preoperative brain herniation for longer than 90 min, and a GCS of 3–5 points) to 47.7% in patients with any three or more of these six risk factors (p<0.001). It was necessary to review carefully the clinical histories of the patients before surgery (the exact time of the head injury, time from head injury to second coma or other obvious neurological deteriorations [based on the first-aid record and/or information provided by the patient's companions], time from the head injury to insensitive pupillary light reflex or even mydriasis [based on the first aid record], and occurrence of hemorrhagic shock) and the neuroradiological imaging, from which the location and volume of the EDH could be obtained.

In this study, we also investigated the aetiology of post-traumatic MCI in patients with severe traumatic EDH, which suggested that the feeding arteries predominantly affected by the EDH-induced gross mechanical compression were the distal (the capillary network and the upstream terminal perforating branches of the cerebral arteries) not the proximal branches (the internal carotid artery and main cerebral arteries, including the anterior/middle/posterior cerebral artery and their affiliated stem segments and perforating branches), as evidenced by the CT angiography (CTA)/magnetic resonance angiography images (data not shown). Incidence of post-traumatic MCI was similar in each of the age intervals and in males and females, which was in line with several other study series.^13,14 Children and teenagers have an equal risk of post-traumatic MCI, compared to adult patients. These data might support a somewhat aggressive, but nonconventional, conservative therapeutic strategy for young patients. A nonconservative therapeutic strategy is more prone to the rationale that decompressive craniectomy would not be adopted until the intracranial hypertension is uncontrolled and the intracerebral infarction changes are significantly massive.^15,16

The duration of preoperative brain herniation is the most prominent contributor to post-traumatic MCI and is significantly higher than the other risk factors. This study demonstrated that bilateral mydriasis independently contributed a 7.08-fold increased risk of traumatic MCI. Patients with preoperative brain herniation lasting longer than 90 min had a more than 6-fold risk of post-traumatic MCI, in contrast to those with herniation lasting less than 60 min. Surgical interventions targeting at hematoma evacuation should be achieved as soon as possible to ameliorate the extent of, and to shorten the duration of, preoperative brain herniation.¹⁷ In our department, every surgically treated patient received a twist-drill drainage before a hematoma-evacuation craniotomy to achieve preoperative decompression quickly, because the preparation for open surgery is time-consuming.

Our results showed that transtemporal EDH had an extremely high risk (16.478-fold; p=0.003) of post-traumatic MCI, compared to nontemporal EDH. The EDH location, especially those in the temporoparietal and temporo-occipital regions, should guide surgical decision making, given the increased risk of MCI they confer. It was suggested that patients suffering from extended preoperative hypotension had a significantly increased risk (13.782-fold; p=0.002) of post-traumatic MCI, which highlighted that fluid resuscitation for sufficient blood supply and cerebral perfusion is extremely important in patients with concurrent hemorrhagic shock.

Many potential clinical indices for predicting post-traumatic MCI remain to be investigated, such as the results of intracranial pressure monitoring, the parameters of CTA/CT venography/CT perfusion images,¹⁰ the core area of the ischemic penumbra,¹⁸ and lesions in the ipsilateral internal carotid artery. Because the results of the above-mentioned risk factors were absent in some of the enrolled patients and some of the risk factors could not be appropriately graded or quantified, we did not subsume these factors into the regression analyses in this study. Future studies are needed to identify the exact roles of these potential risk factors. It would be clinically useful to develop a modified scale on the basis of these risk factors to achieve a convenient bedside evaluation.

This study examined a series of patients who suffered from concurrent acute EDH and brain herniations. The regression analysis revealed that the anatomical location and volume of the EDH, occurrence of preoperative shock, extent and duration of preoperative brain herniation, and GCS scores were the most significant risk factors for post-traumatic MCI. These data suggested that aggressive decompressive craniectomy should be performed in combination with the initial hematoma-evacuation craniotomy in patients with any combination of these risks. Further studies are warranted to evaluate, in an integrated manner, the exact roles of these six currently identified risk factors and other potential risk candidates. The results of these studies will facilitate the consequent development of an effective prediagnostic strategy that is feasible for clinical use and could provide reliable references for surgical decision making.

Footnotes

Acknowledgment

The authors thank the patients who participated in this study.

Author Disclosure Statement

No competing financial interests exist.

References

Liu

, Ma

, Wen

, Shen

, Sheng

, Zhou

, Hu

, Zheng

, and Yang

(2008). Drilling skull plus injection of urokinase in the treatment of epidural haematoma: a preliminary study. Brain. Inj., 22, 199–204.

Nelson

J.A.

(2011). Local skull trephination before transfer is associated with favorable outcomes in cerebral herniation from epidural hematoma. Acad. Emerg. Med., 18, 78–85.

Bullock

M.R.

, Chesnut

, Ghajar

, Gordon

, Hartl

, Newell

D.W.

, Servadei

, Walters

B.C.

, Wilberger

J.E.

; and Surgical Management of Traumatic Brain Injury Author Group. (2006). Surgical management of acute epidural hematomas. Neurosurgery, 58, Suppl. 3, S7–S15; discussion, Si–Siv.

Otani

, Takasato

, Masaoka

, Hayakawa

, Yoshino

, Yatsushige

, Miyawaki

, Sumiyoshi

, Sugawara

, Chikashi

, Takeuchi

, and Suzuki

(2010). Surgical outcome following a decompressive craniectomy for acute epidural hematoma patients presenting with associated massive brain swelling. Acta. Neurochir. Suppl., 106, 261–264.

Mattos

J.P.

, Joaquim

A.F.

, Almeida

J.P.

, Albuquerque

L.A.

, Silva

E.G.

, Marenco

H.A.

, and Oliveira

E.D.

(2010). Decompressive craniectomy in massive cerebral infarction. Arq. Neuropsiquiatr., 68, 339–345.

Mori

, Nakao

, Yamamoto

, and Maeda

(2004). Early external decompressive craniectomy with duroplasty improves functional recovery in patients with massive hemispheric embolic infarction: timing and indication of decompressive surgery for malignant cerebral infarction. Surg. Neurol., 62, 420–429.

Ding

, Yuan

, Guo

, Chen

S.W.

, Gao

W.W.

, Wan,g

, Cao

H.L.

, Ju

S.M.

, Chen

, Zhang

P.Q.

, and Tian

H.L.

(2012). A prospective clinical study of routine repeat computed tomography (CT) after traumatic brain injury (TBI). Brain. Inj., 26, 1211–1216.

Wilson

J.T.

, Slieker

F.J.

, Legrand

, Murray

, Stocchetti

, and Maas

A.I.

(2007). Observer variation in the assessment of outcome in traumatic brain injury: experience from a multicenter, international randomized clinical trial. Neurosurgery, 61, 123–128.

Araujo

J.L.

, Aguiar

U.P.

, Todeschini

A.B.

, Saade

, and Veiga

J.C.

(2012). Epidemiological analysis of 210 cases of surgically treated traumatic extradural hematoma. Rev. Col. Bras. Cir., 39, 268–271.

10.

Mangla

, Ekhom

, Jahromi

B.S.

, Almast

, Mangla

, and Westesson

P.L.

(2014). CT perfusion in acute stroke: know the mimics, potential pitfalls, artifacts, and technical errors. Emerg. Radiol., 21, 49–65.

11.

Subramaniam

, and Hill

M.D.

(2005). Massive cerebral infarction. Neurologist, 11, 150–160.

12.

Manawadu

, Quateen

, and Findlay

J.M.

(2008) Hemicraniectomy for massive middle cerebral artery infarction: a review. Can. J. Neurol. Sci., 35, 544–550.

13.

Irie

, Le, Brocque

, Kenardy

, Bellamy

, Tetsworth

, and Pollard

(2011). Epidemiology of traumatic epidural hematoma in young age. J. Trauma, 71, 847–853.

14.

Tian

H.L.

, Geng

, Cui

Y.H.

, Hu

, Xu

, Cao

H.L.

, Chen

S.W.

, and Chen

(2008). Risk factors for posttraumatic cerebral infarction in patients with moderate or severe head trauma. Neurosurg. Rev., 31, 431–436.

15.

Lau

, El-Sayed

A.M.

, Ziewacz

J.E.

, Jayachandran

, Huq

F.S.

, Zamora-Berridi

G.J.

, Davis

M.C.

, and Sullivan

S.E.

(2012). Postoperative outcomes following closed head injury and craniotomy for evacuation of hematoma in patients older than 80 years. J. Neurosurg., 116, 234–245.

16.

Haddad

S.H.

, and Arabi

Y.M.

(2012). Critical care management of severe traumatic brain injury in adults. Scand. J. Trauma. Resusc. Emerg. Med., 20, 12.

17.

Smith

S.W.

, Clark

, Nelson

, Heegaard

, Lufkin

K.C.

, and Ruiz

(2010) Emergency department skull trephination for epidural hematoma in patients who are awake but deteriorate rapidly. J. Emerg. Med., 39, 377–383.

18.

Kruetzelmann

, Hartmann

, Beck

, Juettler

, Singer

O.C.

, Köhrmann

, Kersten

J.F.

, Sobesky

, Gerloff

, Villringer

, Fiehler

, Neumann-Haefelin

, Schellinger

P.D.

, Röther

, Thomalla

; and Clinical Trial Net of the German Competence Network Stroke. (2014). Combining magnetic resonance imaging within six-hours of symptom onset with clinical follow-up at 24 h improves prediction of ‘malignant’ middle cerebral artery infarction. Int. J. Stroke, 9, 210–214.