Do Age and Anticoagulants Affect the Natural History of Acute Subdural Hematomas?

Abstract

Acute subdural hematoma is a serious complication following traumatic brain injury. Large volume hematomas or those with underlying brain injury can cause mass effect, midline shift, and eventually herniation of the brain. Acute subdural hematomas in the young are associated with high-energy trauma and often have underlying contusions, while acute subdural hematomas in the elderly are associated with minor trauma and an absence of underlying contusions, even though the elderly are more likely to be on anticoagulants or anti-platelet therapy. In the young patients with high impact injuries the hematomas tend to be small and the underlying brain injury and swelling is responsible for the increased intracranial pressure and midline shift. In the elderly, the injuries are low impact (e.g fall from standing), the underlying brain is intact, and the volume of the hematoma itself produces symptoms. In addition the use of anticoagulants and antiplatelet agents in the elderly population has been thought to be a poor prognostic indicator and is considered to be responsible for larger hematomas and poor outcome. When managed conservatively, acute subdural hematomas can sometimes progress to chronic subdural hematoma formation, further enlargement, seizures, and progressive midline shift. Another potential difference in the young and the elderly is brain atrophy, which increases the potential space to accommodate a larger hematoma. It is not known if these two groups differ in other ways that might have implications for treatment or prognosis. In this paper, we investigate the clinical course of 80 patients admitted to our institution with acute subdural hematomas, to identify differences in patients above or below the age of 65 years. The natural progression/resolution of acute subdural hematomas was mapped by measuring volume expansion/regression over time. In this retrospective chart review, we investigated clinical baseline metrics and subsequent volumetric expansion outcomes between patients < 65 years old (N=44) and those > 65 years old (N=36). Volume was estimated by the ABC/2 method. We observed a statistically significant difference between groups in use of anticoagulants χ² =40.305 with p < 0.001, corrective platelet administration χ² =19.380 with p < 0.001, gender χ² =14.573 with p < 0.001, and Glasgow Coma Scale with χ² =23.125 (p=0.026). Overall outcomes were similar in the two groups. Younger patients on average had worse presenting GCS scores, but recovered comparable to older patients. No significant difference in rate of volume expansion, resolution time, or need for surgical treatment was seen between these two groups. We conclude that the initial volume, size, and severity of subdural hematoma determined by the Glasgow Coma Scale score is more likely to predict surgery or future expansion than age of the patient. Patients on oral anti-coagulants that are given appropriate medical reversal agents early do quite well and no impact on the eventual outcome could be demonstrated. Further work is needed to establish better predictors of future volume expansion, and progression to chronic subdural hematoma based on improved severity scales.

INTRODUCTION

A common complication of traumatic brain injury (TBI) is subdural hematoma. ~22% of TBI patients can present to the emergency department with subdural hematoma, which contributes to the extensive morbidity and mortality associated with TBI [1,2]. Common complications of subdural hematomas include mass effect resulting in altered cerebral blood flow and increase intracranial pressure [3]. Furthermore, subdural hematomas result in an elevated risk of seizure disorders. The causes of subdural hematomas are variable and consist of falls, particularly in the elderly, assault, other forms of trauma, and child abuse [4–6]. Notably, use of anticoagulants or antiplatelet agents may predispose to subdural hematomas or potentially worsen outcome. Management of subdural hematomas is varied, frequently employing either medical or surgical approaches [7]. The indications for surgery include a hematoma more than 10 mm thick and/or a midline shift of 5 mm or more. In addition any hematoma with a GCS score of 8 or below or an ICP of more than 20 mmHg is surgically explored. The size indication is more relevant for the elderly where the underlying brain is intact and the cause is usually bleeding from a cortical bleeder. The Glasgow Coma Score is more important for the young with high impact injuries and underlying brain injury. When patients are managed conservatively, the incidence of progression to chronic subdural hematoma is not fully known [8]. Ahmed and colleagues found that 5 out of 27 patients with acute subdural hematoma progressed to chronic subdural hematoma [9]. Patients on anticoagulants prior to admission, however, were excluded from their study. In some studies anti-coagulation therapy has been shown to increase mortality from large acute subdural hematomas and limit spontaneous resolution [10]. Additionally, elderly patients on anti-coagulant therapy are more likely to get acute subdural hematomas if they fall and receive head trauma [11]. It has been suggested that resolution of acute subdural hematomas may be more common for patients not on anti-coagulants [12]. Resolution is also thought to be dependent on the size of the initial subdural hematoma [13].

If large acute subdural hematomas are not managed early, the volume can expand acutely, leading to progressive midline shift and possible brain herniation [14]. Often these subdural hematomas have an underlying cortical bleeder, which needs to be controlled. The risk for rapid deterioration is highest within the first 4 hours after severe head trauma [15]. Large acute subdural hematomas are more likely to require surgical intervention [16]. Poorly managed subdural hematomas can lead to spreading cortical depression as well, which is a risk factor for seizure [17]. Okumura and colleagues found that 10 out of 96 patients progressed from mild acute subdural hematoma to chronic subdural hematoma. The mean age of the patients who had progressive subdural hematoma was 63.1 years [18]. Severity of initial injury may account for worse outcome over time [19]. Magnetic resonance imaging has been useful in monitoring subdural hematoma expansion or resolution [20,21]. ABC/2 volumetric calculation with computed tomography images has been verified as an accurate estimation approach for measuring subdural progression over time [22]. In this study, we look at the natural course of subdural hematomas over time using the ABC/2 volumetric calculation comparing young (<65 years old) and older (>65 years old) patients as well as by use of anti-coagulant therapy. The age range for young patients was 23–64 years old where the age range for the older patients was 66–91 years old.

METHODS

The West Virginia University Institutional Review Board approved the protocol for this study, and the study was performed in accordance with The Code of Ethics of the World Medical Association. We collected retrospective data on 80 patients with acute subdural hematoma that presented to Ruby Memorial Hospital from 2010 to 2014 and were managed by the Department of Neurosurgery. For this type of study, formal consent is not required. Statistical consultation and a power analysis revealed that 30 patients were needed per group to determine statistical significance. Only patients with a CT scan at presentation, confirming acute SDH, as well as follow-up imaging between 3 and 6 weeks, were included for analysis. Data collected included: patient age, sex, presence of anti-coagulant or anti-platelet therapy on admission, Glasgow coma scale score at presentation, and surgical or procedural intervention. Other important variables gathered were presenting symptoms, laterality of convexity subdural hematoma, medical interventions given, midline shift, and clinical values for prothrombin time and international normalized ratio. The cohort was divided into those < 65 years old (N=44) and those >65 years old (N=36). Subdural volume was calculated at presentation and at a 1-month follow-up visit using AB(C^*0.5)/2 (Figure 1). A is the width in cm, B is the length in cm, and C is the number of slices. ABC result was multiplied by 0.5 to account for the predetermined slice imaging thickness of 0.5 cm. The entire multiplied product was divided by 2 to account for the crescent volumetric distribution of the subdural hematoma. Volume progression was grouped into four categories: 0–30% of original = resolution, 30–65% of original = reduction, 65–100% of original = stasis, and > 100% = expansion (Figure 2). Volume progression was compared based on age category and separately according to anti-coagulant use. We furthermore excluded patients requiring surgery from the volumetric measurements. 34 conservatively managed patients < 65 years old and 27 conservatively managed patients > 65 years old had scans that were evaluated for volume expansion.

Figure 1.

Calculating Volume. Volume is calculated by ABC*5/2 where A is the width in cm, B is the length in cm, and C is the number of slices where the subdural hematoma is present. ABC is multiplied by 5 based on the predetermined set thickness of the CT slices. The total value is divided by 2 to account for the crescent volumetric shape of the subdural hematoma.

Figure 2.

Volume Progression. Representative images showing examples of volume resolution, reduction, stasis, and expansion.

Statistical analysis

All statistical analysis was performed using JMP 9.0. Categorical variables were analyzed by χ² analysis. Dichotomous variables were analyzed by a Fischer’s exact test with a 2-tailed distribution. A t-test was used to compare clinical test values between groups. Mean ± standard error is reported with a mean difference ± standard error of the mean. P < 0.05 is considered statistically significant for all groups.

RESULTS

Metrics by age

A significant difference in gender was observed between patients > 65 years old and < 65 years old χ² =14.573 with p < 0.001. Of those patients > 65 years old, 69% were female, whereas only 27% were female in those patients < 65 years old. No significant differences between groups were observed for the five reported symptoms: headache, altered mental status, paresis, seizure, and none. A significant difference in mechanism of injury was observed between patients > 65 years old and < 65 years old χ² =31.946 with p < 0.001. Minor falls, i.e. from standing, accounted for 80% of the subdural hematomas in those > 65 years old, whereas only 41% of subdural hematomas were from falls in those < 65 years old. Motor vehicle collision (41%) and assault (7%) caused a significant proportion of subdural hematomas in those < 65 years old, but were minimal contributors for subdural hematomas in those > 65 years old (6% and 2% respectively). The laterality of subdural hematoma was not statistically significant between groups (Table 1).

Table 1. Metrics by Age.

Gender χ2 =14.573 P < 0.001***	Male	Female
Younger (< 65 years old)	N=32 (73%)	N=12 (27%)
Older (> 65 years old)	N=11 (31%)	N=25 (69%)

Symptoms χ2 =20.859 p=0.1414	Altered Mental Status	Headache	None	Paresis	Seizure
Younger (< 65 years old)	N=13 (30%)	N=18 (41%)	N=2 (4%)	N=4 (9%)	N=7 (16%)
Older (> 65 years old)	N=10 (28%)	N=13 (36%)	N=7 (20%)	N=3 (8%)	N=3 (8%)

Mechanism χ2 =31.946 p<0.01**	Fall	Motor Vehicle Collision	Hypertensive	Assault	Unknown
Younger (< 65 years old)	N=18 (41%)	N=18 (41%)	N=0 (0%)	N=3 (7%)	N=5 (11%)
Older (> 65 years old)	N=29 (80%)	N=2 (6%)	N=1 (6%)	N=2 (2%)	N=2 (6%)

Convexity SDH Laterality χ2 =19.501 p=0.1919	Left	Right	Both
Younger (< 65 years old)	N=19 (43%)	N=18 (41%)	N=7 (16%)
Older (> 65 2years old)	N=15 (42%)	N=14 (39%)	N=7 (19%)

Clinical variables by age

A significant difference in the use of anti-coagulants was seen between those patients > 65 years old and those < 65 years old, χ² =40.305 with p < 0.001. In those < 65 years old, 87% were not using any anti-coagulant or antiplatelet medication on admission. In those > 65 years old, 39% were on aspirin, 17% were on aspirin + clopidogrel, and 8% were on aspirin + warfarin. A significant difference was also seen between groups for the medical interventions: prothrombin complex concentrate and platelet infusion. 5 patients > 65 years old were given prothrombin complex concentrate, where only 1 patient < 65 years old was given this intervention (Fischer’s Exact test P < 0.01). Patients given prothrombin complex concentrate (PCC) had a −.33 correlation with time to complete subdural resolution (95% CI −.571 to −0.037). This indicates faster recovery when PCC was given. 19 patients > 65 years old were given platelets, where 4 patients < 65 years old were given platelets (Fisher’s Exact test P < 0.001). No significant difference between groups was observed for percentages treated surgically, or type of surgery (Table 2). Patients requiring surgery were not analyzed in the serial volume expansion/regression measurements. Only patients managed conservatively were evaluated for volume expansion/regression.

Table 2. Clinical Variables by Age.

Anticoagulation χ2 =40.305 P < 0.001***	Aspirin	Warfarin	Aspirin + Clopidogrel	Aspirin + Warfarin	Aspirin + Clopid. + Warfarin	Other	None
Younger (< 65 years old)	N=1 (2%)	N=0 (0%)	N=1 (2%)	N=3 (7%)	N=0 (0%)	N=1 (2%)	N=38 (87%)
Older (> 65 years old)	N=14 (39%)	N=1 (3%)	N=6 (17%)	N=3 (8%)	N=2 (6%)	N=0 (0%)	N=10 (27%)

Medical Intervention	Fresh Frozen Plasma	Prothrombin Complex Concentrate	Vitamin K	Platelets
Younger (< 65 years old)	N=3 Yes N=41 No	N=1 Yes N=43 No	N=3 Yes N=41 No	N=4 Yes N=40 No
Older (> 65 years old)	N=4 Yes N=32 No	N=5 Yes N=31 No	N=5 Yes N=31 No	N=19 Yes N=17 No
	Fischer’s Exact Test:2-tail difference between groups p=0.69	Fischer’s Exact Test:2-tail difference between groups p<0.01*	Fischer’s Exact Test:2-tail difference between groups p=0.45	Fischer’s Exact Test:2-tail difference between groups p< 0.001***

Surgery Required Fischer’s Exact Test:2-tail difference between groups p=1	Yes	No
Younger (< 65 years old)	N=10	N=34
Older (> 65 years old)	N=9	N=27

Type of Surgery χ2 =4.226 p=0.238	Bur-Hole Drainage	Craniotomy	Twist-drill Craniostomy
Younger (< 65 years old)	N=1 (11%)	N=8 (89%)	N=0 (0%)
Older (> 65 years old)	N=0 (0%)	N=7 (78%)	N=2 (22%)

Clinical tests by age

No significant difference between groups was observed for prothrombin time −0.99 ± 1.01 (p=0.16). Those > 65 years old had a mean of 12.64 ± 0.79 compared to 11.65 ± 0.63 for those < 65 years old. No significant difference between groups was also found for the international normalized ratio −0.15 ± 0.12 (p=0.11). Those > 65 years old had a mean of 1.28 ± 0.09 compared to 1.13 ± 0.07 for those < 65 years old. Presenting Glasgow Coma Scale Scores were significantly different −2.68 ± 0.74 (p < 0.001). Those > 65 years old had a mean of 14.24 ± 0.6 compared to 11.56 0.54 for those < 65 years old. No significant difference in initial subdural hematoma volume was observed across age groups, despite the difference in other clinical variables (Table 3). The mean volume in cm³ ± SE for the younger group is 30.3 ± 15.08 and for the older group 13.99 ± 2.27. The standard deviation for the younger group is 100.02 and for the older group 13.6.

Table 3. Clinical Tests by Age.

Clinical Tests	Prothrombin Time	International Normalized Ratio	Glascow Coma Scale	Initial Subdural Hematoma Volume
Younger (< 65 years old)	Mean: 11.65 SE: 0.63	Mean: 1.13 SE: 0.07	Mean: 11.56 SE: 0.54	Mean: 30.3 SE: 15.08
Older (> 65 years old)	Mean: 12.64 SE: 0.79	Mean: 1.28 SE: 0.09	Mean: 14.24 SE: 0.60	Mean 13.99 SE: 2.27
	Mean Difference: −0.99 ± 1.01 t=0.98 p=0.16	Mean Difference: −0.15 ± 0.12 t=1.25 p=0.11	Mean Difference: −2.68 ± 0.74 t=3.61 p < 0.001*	Mean Difference: 16.32 ± 15.25 t=1.07 p=0.855

Outcomes

No patients suffered acute mortality following subdural hematoma. No significant difference between groups was observed on volume change at the 1-month follow up (χ² =1.514, p=0.68). In those > 65 years old, 40.74% had resolution, 14, 81% had reduction, 14.81% had stasis, and 29.63% had expansion. Of those patients that had expansion, 75% progressed to chronic subdural hematoma. In those < 65 years old, 52.94% had resolution, 17.65% had reduction, 8.82% had stasis, and 20.59% had expansion. Of those patients that had expansion, 71% progressed to chronic subdural hematoma. No significant differences between groups were seen in midline shift (mean difference 0.101 ± 0.19, p=0.703). Mean size difference of subdural hematoma at baseline and the 1-month follow up appointment were also not statistically significant between groups. We also compared volume changes according to the use of anti-coagulants. No significant differences were seen between groups. Of patients with no anti-coagulant use, 56.41% had resolution, 12.82% had reduction, 7.69% had stasis, and 23.08% had expansion. For those on aspirin, 23.08% had resolution, 23.08% had reduction, 15.38% had reduction, and 38.46% had expansion. For those on aspirin + clopidogrel, 50% had resolution, 0% had reduction, 50% had stasis, and 0% had expansion. For those on aspirin + warfarin, 33.33% had resolution, 33.33% had reduction, 0% had stasis, and 33.33% had expansion (Table 4). When patients that required surgery were excluded, the χ² statistic was 0.51.

Table 4. Outcomes.

Volume Change by Age χ2 =1.514 p=0.68	Resolution	Reduction	Stasis	Expansion
Younger (< 65 years old)	N=18 (52.94%)	N=6 (17.65%)	N=3 (8.82%)	N=7 (20.59%)
Older (> 65 years old)	N=11 (40.74%)	N=4 (14.81%)	N=4 (14.81%)	N=8 (29.63%)

Volume Change by Anticoagulant χ2 =20.17 p=0.51	Resolution	Reduction	Stasis	Expansion
Aspirin	N=3 (23.08%)	N=3 (23.08%)	N=2 (15.38%)	N=5 (38.46%)
Plavix	N=1 (100%)	N=0 (0%)	N=0 (0%)	N=0 (0%)
Warfarin	N=0 (0%)	N=0 (0%)	N=1 (100%)	N=0 (0%)
Aspirin + Plavix	N=1 (50%)	N=0 (0%)	N=1 (50%)	N=0 (0%)
Aspirin + Warfarin	N=1 (33.33%)	N=1 (33.33%)	N=0 (0%)	N=1 (33.33%)
Aspirin + Plavix + Warfarin	N=1 (100%)	N=0 (0%)	N=0 (0%)	N=0 (0%)
Other	N=0 (0%)	N=1 (100%)	N=0 (0%)	N=0 (0%)
None	N=22 (56.41%)	N=5 (12.82%)	N=3 (7.69%)	N=9 (23.08%)

DISCUSSION

Patients with subdural hematoma and underlying comorbidities often require advanced care from tertiary care centers. Most of these patients can be successfully managed conservatively, but a subset requires surgical intervention [19]. Surgical intervention is often necessary when the subdural hematoma is large, expands over time, or is associated with a significant midline shift [14]. Appropriate clinical evaluation is critical for determining when to manage conservatively vs. surgically in these cases [23]. In our cohort, 61 patients were managed conservatively and 19 required surgery for chronic subdural hematoma. The surgical procedures utilized for hematoma evacuation were Bur-hole drainage, craniotomy, or twist drill craniotomy. Craniotomy and evacuation of acute subdural hematoma was performed early in the admission for acute subdural hematomas. For chronic enlarging hematomas either a twist drill craniostomy (bedside placement of a subdural drain with a 5 mm drill to access the hematoma) or a burr-hole drainage (one or more14 mm opening/openings in the skull to allow washout of the hematoma and placement of a drain) in the operating room, were utilized.

Ahmed and colleagues reported that 19% of patients managed conservatively progress to chronic subdural hematoma [9]. We show that 25% of patients in our study had volume expansion at a 1-month follow-up visit with 18.25% progressing to chronic subdural hematoma. We sought to investigate the natural progression of conservatively managed subdural hematomas in a younger cohort (< 65 years old) and an older cohort (> 65 years old). We hypothesized that older patients would have a greater likelihood for subdural hematoma expansion.

We found that volume changes however were not statistically different between the two groups. Despite having a statistically significant difference in the mechanisms of injury, the initial volume and volume changes were similar between groups. Surgical intervention and type of surgery were also not statistically significant between groups. The younger cohort of patients had a mean Glasgow Coma Scale score of 11.56 compared to 14.24 in the older cohort. This suggests that severity of injury may have been higher in younger cohort patients, especially considering that the older cohort patients had a statistically significant increase in use of anticoagulants and medical interventions. The mechanisms of injury suggest that the younger cohort had more motor vehicle collisions and assault (high energy trauma), where the older cohort had minor falls (low energy trauma).

The use of anti-coagulants may account for why falls were higher in the older cohort. If medical reversal interventions are given early to counteract the oral anticoagulants, patients will often do quite well [10]. Recent evidence suggests that anticoagulant use in those > 65 years old with small subdural hematomas is not a significant determinant of volume expansion [24]. Rapid resolution of subdural hematoma across age groups may be more common than previously thought [12]. Case series have reported rapid resolution of subdural hematomas in young patients [25], but we did not observe rapid resolution in any of our cases. It has been noted that the severity of the traumatic brain injury causing the subdural hematoma is more important than the age of the patient in determining the susceptibility for delayed deterioration [26]. Gerard and colleagues suggest that prothrombin time and the international normalized ratio should be quickly evaluated to determine the necessity of reversal agents and medical management [27]. If anticoagulants are not reversed in subdural hematomas quickly, they can accentuate the severity of injury and bleeding duration [28]. Our data shows that when reversal agents are given rapidly and the coagulation defect corrected, there were no significant adverse effects of anticoagulant therapy for subdural hematoma.

It has been hypothesized that an increase in intracerebral pressure following a tear of a cortical bridging vein accounts for subdural hematoma expansion [29]. The increase in intracerebral pressure elevates cerebral venous pressure thereby causing an increased risk for re-bleeding and higher levels of morbidity [30]. If the re-bleeding is associated with a potentially increased subdural space, then acute subdural hematomas can progress to chronic subdural hematomas [31]. Most subdural hematomas however spontaneously resolve with pre-existing subdural hygromas being more likely to progress to chronic subdural hematoma. Brain atrophy in the elderly may also allow stasis of the acute hematoma and favor conversion to a chronic hematoma. Our data shows that 55.55% of patients > 65 years old and 68.59% of patients < 65 years old had reduction or resolution of the subdural hematoma at 1 month when managed conservatively, while 14.91% of the younger cohort and 22.22% of the older cohort went on to develop chronic subdural hematoma. This supports the finding of older patients with pre-existing brain atrophy and enlarged spaces as being more likely to develop chronic subdural hematomas, but the presence of an acute subdural hematoma does not always imply progression to chronic subdural hematoma.

CONCLUSION

We failed to observe a statistically significant difference between groups for progression of acute subdural hematoma to chronic subdural hematomas based on age. The use of surgical intervention was also not statistically significant between groups. The severity of initial injury measured by GCS is more likely to affect the overall outcome with our data showing young patients are more likely to develop subdural hematoma following high energy trauma and older patients often present after minor falls. Among conservatively managed acute subdural hematomas, the risk of conversion to chronic SDH was 14.91% in the younger cohort and 22.22% in the older cohort. Volume expansion may depend on the mechanism and severity of injury correlated with the age of the patient. Although the younger cohort had worse GCS scores at admission, the resolution of subdural hematomas was comparable to the older cohort with better GCS scores. The use of anti-coagulants should be determined early and treated rapidly with reversal agents to prevent expansion of subdural hematoma. Appropriate management also includes the use of imaging to determine volume expansion, careful clinical exams, and the use of surgical intervention where warranted. Further work is needed to establish a better clinical scale for the evaluation of subdural hematoma severity. The progression to chronic subdural hematoma is not uncommon, ~18–19%, but must be considered, diagnosed and managed appropriately in order to reduce overall morbidity and mortality.

ABBREVIATIONS

CT

Computed Tomography

FFP

Fresh Frozen Plasma

GCS

Glasgow Coma Scale

INR

International Normalized Ration

PCC

Prothrombin Complex Concentrate

PT

Prothrombin Time

TBI

Traumatic Brain Injury