Acute subdural haematoma
1 Management of subdural haematoma, including appropriate timing of operation, and technique.
2F01
A 70-year-old, 85 kg, and 1.83 m tall man presented to hospital with a 5-day history of frontal headache. Previous medical history included well-controlled AF, migraines, gastro-oesophageal reflux disease, and osteoarthritis. He did not recall any history of trauma. Neurosurgical services in the hospital were contacted.
His medication included: warfarin, digoxin 125 micrograms/day, diltiazem 300 mg/day, pantoprazole 20 mg/day, Seretide® and salbutmol inhalers.
On admission, examination findings were: BP 160/100, pulse ventricular rate 81, no heart murmurs, chest clear. GCS 15. No focal deficits.
Initial investigations were:
◆ Na+ 136 mmol/L, K+ 4.9 mmol/L, urea 6.4 mmol/L, Cr 63 micromoles/L, estimated glomerular filtration rate (eGFR) >60 mL/min
◆ Hb 143 g/L, WCC 13.9 × 109/L, Plt 142 × 109/L
◆ Clotting: APTT 29 s (ratio 1.0), INR 2.7, fibrinogen 3.9 g/L
◆ Brain CT scan: 11 mm left-sided subdural haematoma (see Figure 7.2).
Fig. 7.2 Admission axial non-contrasted CT scan showing an acute left subdural haematoma, 11 mm in depth with a 4 mm midline shift.
The generally accepted anatomical source of bleeding in subdural haemorrhage is from the bridging veins passing from the cortical surface to the large venous sinuses, most particularly the superior sagittal sinus. These haemorrhages are rather more common in the elderly. This is because, as one gets older, the brain shrinks and the bridging veins have longer to travel and are therefore more liable to being torn by trauma. This can be quite trivial and not remembered by the patient. Often the patient is asymptomatic for some time, but, as blood products break down, the haematoma tends to enlarge and cause more mass effect. In this case, the patient’s risk of an intracranial haemorrhage is increased by him being on warfarin therapy. This is quite a frequent scenario in neurosurgery and requires careful management from anaesthetics, neurosurgery, and the haematology service.
In general, an acute intracranial haemorrhage will always require normalization of the INR. The length of time off anticoagulants is a trade-off between the risks of a rebleed and thrombosis, which will have to be judged on an individual basis.
Consideration has to be given as to why a patient was on warfarin, and, if an additional specialty were to have started warfarin (such as cardiac surgery, cardiology, or neurology), then they should be included in this discussion. The patient may be taking warfarin for a recently diagnosed thrombotic event or because they have a metallic heart valve. Particularly in patients who have metallic heart valves, there is a fine balance to be struck between the safe reversal of their clotting and a thrombotic event.
In this patient’s case, it was to reduce the chance of thrombotic events secondary to AF. The decision was made to reverse the prolonged INR with 5 mg of IV vitamin K and 2000 U of Beriplex® (a form of prothrombin complex concentrate (PCC), a combination of blood clotting factors II, VII, IX, and X, as well as protein C and S). This corrected the INR to 1.1.
The patient was admitted to the neurosurgical ward. It was considered most appropriate to keep him under observation and defer surgery. The evacuation of an acute subdural haematoma (ASDH) (when the appearance is white on CT scan) requires a craniotomy, as the blood is thick and firm. Over 1–2 weeks, the subdural blood will liquefy (turning black on CT) and may be evacuated via one or two small burr holes (14 mm in diameter). This is much less traumatic for the patient. Given that his only symptom was headache, he had no focal deficits, and he remained fully alert and oriented, it was decided to wait.
As the haematoma will go through various phases (from liquid to solid to liquid again), it was felt best to observe the patient until the haematoma had liquefied, then to evacuate, unless his condition deteriorated. He was noticed to be rather sleepy and slightly confused on the eighth day after admission, reducing his GCS from 15 to 14. A repeat CT scan was performed, showing a new midline shift of >1 cm (see Figure 7.3).
Fig. 7.3 Axial non-contrasted CT scan showing a left mixed-density subdural haematoma, now with a 13 mm midline shift and effacement of the left lateral ventricle, prior to surgery.
Given the deterioration in GCS and the worsening appearance on CT scan, the decision was made to evacuate the haematoma now. The appearance of the subdural on CT was still dense (white) and so was not yet amenable to surgery via burr hole; therefore, he was prepared for a craniotomy.
Classically, a subdural haematoma occurs due to venous bleeding, and an extradural haematoma (referred to as an epidural haematoma in North American literature) from arterial bleeding (see Table 7.1 and compare Figures 7.2 and 7.4 which contrasts the patient’s initial CT head scan with another CT scan of a patient demonstrating an extradural haematoma).
Table 7.1 A comparison of common features of subdural and extradural haematomas
Subdural haematoma | Extradural haematoma | |
---|---|---|
Source |
Bridging veins |
Middle meningeal artery, commonly from fracture of the overlying squamous temporal bone (can be caused by any meningeal artery), or from ooze from the bone at any fracture site |
Shape |
Crescentic |
Lentiform |
Relationship to skull suture lines |
Will spread over the whole hemisphere |
Classically does not cross sutures due to adherence of periostium |
Acute/chronic |
Can be either |
Almost always acute |
Urgency of treatment |
If chronic and causing minimal symptoms, may be managed expectantly. If acute, especially in the young, and with mass effect, should be treated as per extradural haematoma |
Generally, this is a lifesaving emergency of the highest urgency |
Age of patient |
Often older, although acute subdural haematoma in a young person indicates a very significant injury |
Usually younger |
Fig. 7.4 Example of an extradural haematoma. Note the lentiform shape (convex), compared to the concave subdural haematomas. Also note the marked midline shift.
Local anaesthesia is sometimes adopted for burr hole surgery if the patient is compliant. It is also sometimes used when resecting tumours in eloquent areas. For trauma craniotomies, however, there is no rationale to favour awake surgery, as the urgency of the situation, as well as the need for surgical control, dictates the requirement for general anaesthesia.
We stress that this decision should be made before embarking on the procedure. We do not think it is good practice to start under local anaesthesia and to then convert to a general anaesthetic mid-procedure (potentially with the head open).
There are two broad categories of technique used for an asleep technique in neuroanaesthesia: TIVA and a volatile-based technique.
Propofol theoretically has fewer deleterious effects on autoregulation, reduces CMRO2, and thus optimizes the cerebral blood flow, and consequently the ICP. Propofol has an antiemetic profile, reducing the likelihood of PONV. In the presence of a critically elevated ICP, arguably it is the technique of choice.
However, due to its accumulation in adipose tissue, propofol may lead to prolonged wakening. As end-tidal propofol concentrations cannot be routinely measured, a depth of anaesthesia monitor is advisable. One may wish to continue a propofol infusion when receiving an intubated head injury transfer (particularly as they are most likely to have been established on a propofol infusion for the transfer).
All volatile agents reduce cerebral autoregulation to a varying degree. Sevoflurane starts to affect autoregulation at 1.5 MAC; isoflurane starts to effect autoregulation at 1.0 MAC, and desflurane at only 0.5 MAC.
The counter argument is that, when the cranium is actually open, the ICP is at atmospheric pressure anyway, so does this matter as much? The co-administration of an agent, such as remifentanil, reduces this deleterious effect.
There are surprisingly few direct comparisons between sevoflurane and desflurane on longer-duration intracranial cases. One of the few studies, which directly compares sevoflurane against desflurane, found no difference in emergence time (12.2 ± 4.9 min with sevoflurane; 10.8 ± 7.2 min with desflurane), but a marginally quicker time to extubation (15.2 ± 3.0 min for sevoflurane; 11.3 ± 3.9 min for desflurane; p <0.001) and recovery (18.2 ± 2.3 for sevoflurance; 12.4 ± 7.7 for desflurance; p <0.001). Results shown are means with standard deviations.
The short orientation memory concentration test (SOMCT) score, as a test for the return of mental capacity, was marginally better with desflurane; however, the scores with sevoflurane were still within the normal range.
Worries about desflurane’s effect on autoregulation may be more theoretical than real. A pragmatic use may be when there is a definite advantage such as an elevated BMI.
A 20 g left radial arterial line and BIS depth of anaesthesia monitor were placed prior to induction. Anaesthesia was induced with propofol 150 mg and rocuronium 80 mg, and the hypertensive response to intubation obtunded with a remifentanil infusion, supplemented with a bolus of remifentanil 70 micrograms. The trachea was intubated with an 8.0 mm internal diameter armoured ETT. The maintenance of anaesthesia was with an oxygen/air/sevoflurane technique, with continuation of the remifentanil infusion (basic rate 0.1 micrograms/kg/min), and the target for the BIS monitor was 40–45. A urinary catheter was sited after induction. Intermittent metaraminol boluses were used to maintain an adequate CPP. The patient was ventilated to maintain ETCO2 in the low normocapnia range.
Analgesia was with IV paracetamol and fentanyl 50 micrograms administered towards the end of the operation, to supplement the 20 mL 50/50 mix of 1% lidocaine with 1/200 000 adrenaline/0.5% bupivacaine infiltrated by the surgeon. Granisetron was administered as an antiemetic.
The patient woke up with a GCS of 14 (due to confusion). Post-operative analgesia was with regular paracetamol, supplemented with PCA morphine and dihydrocodeine, as required.
On the day after the procedure, his GCS fell from 13 (E3 V4 M6) to 11 (E4 V2 M5), and he appeared vacant. He then had two episodes of facial twitching, lasting for 2 and then 3 min. A CT scan of the head performed post-operatively is shown in Figure 7.5.
Fig. 7.5 Post-operative scan showing an almost complete removal of a subdural collection, some expected post-operative intracranial air (pneumocephalus), and resolution of the midline shift.
He was given a loading dose of 1 g phenytoin, followed by supplementary levetiracetam.
His seizure activity ceased. However, his INR, which had originally been corrected to within the normal range, was prolonged 3 days later to 1.8 and then 2.1.
A drug interaction between the newly commenced anticonvulsant and any residual warfarin still in the body (by displacement of protein binding) was considered. This would have been problematic, as the anticonvulsant was required to stop seizure activity.
However, on discussion with haematology colleagues, it was felt that the prolonged INR was due to a deficiency of vitamin K. Supplements of vitamin K were administered, and the INR returned to normal.
There were no further seizures. After a few more days on the ward, the patient was discharged home with an appropriate support package. GCS on discharge was 15, and activities of daily living had returned to normal. He was scheduled for follow-up in the outpatient clinic.
Summary
The evacuation of a subdural haematoma represents a significant caseload in emergency neurosurgery. There are differences between subdural and extradural haematomas. Of note, subdural haematomas may go through various stages of maturation, and the timing of any intervention may take this into account, as well as any correction of clotting abnormality. As patients may have other comorbidities, there may be discussions with other specialties, especially if the patient has been commenced on anticoagulants by these colleagues.