This is the third awake case in the curriculum and the first where the choice of drug is dictated by which nucleus the surgeon is aiming at. It is also the only one where you deliberately withdraw the patient's medication to make them worse, then bolt their head to the table — and the escape route from that is something you have to arrange before you start.
The target chooses the anaesthetic. Globus pallidus interna neurons receive predominantly GABAergic input from the putamen and external pallidum, so they are far more suppressed by propofol than subthalamic neurons are — subthalamic recordings survive low doses of sevoflurane, desflurane, propofol, ketamine or dexmedetomidine, while pallidal recordings are often impossible under general anaesthesia, and in dystonia pallidal firing rates fall substantially under propofol compared with no sedation. Non-GABAergic agents and short-acting opioids have minimal effect, which is why dexmedetomidine below 0.6 µg/kg/h with a BIS above 80 has become the preferred sedative.
Three days in this curriculum have an awake patient in them, and it is worth saying plainly what makes each different, because they are easy to blur together.
Neuro 2 day 04 was the awake craniotomy, where the question was which phase actually needs the patient awake. Day 10 was epilepsy surgery, where the danger was that your drug could manufacture the finding. This day is deep brain stimulation, and its organising fact is different again: which drugs you may use is determined by which nucleus the electrode is heading for.
Deep brain stimulation places electrodes into deep structures, records from them to confirm position, stimulates to test the clinical effect, and then connects them to a pulse generator implanted below the clavicle or in the abdomen. The system is three parts: intracranial electrodes, the generator in its titanium housing, and the extension cable between them.
The target follows the indication:
| Indication | Target |
|---|---|
| Parkinson’s disease | Subthalamic nucleus (also pallidum) |
| Dystonia | Globus pallidus interna |
| Essential tremor | Thalamic Vim |
| Obsessive-compulsive disorder | Anterior limb of internal capsule, subthalamic nucleus |
| Epilepsy | Anterior nucleus of the thalamus |
| Addiction | Nucleus accumbens |
| Alzheimer’s disease | Fornix |
Those last three signal how fast the field is moving. What matters for you is the first two, and the difference between them.
GABAergic anaesthetic agents affect both background activity and neuronal spike activity, and — this is the key sentence — the effects are not homogeneous across the brain.
Globus pallidus interna neurons receive predominantly GABAergic input from the putamen and the external pallidum. Give propofol, a GABAergic drug, and you are adding to an input those neurons are already dominated by. They fall quiet.
Subthalamic neurons do not have that input profile, so they are comparatively resistant.
The consequences follow without needing to be memorised separately:
Hence dexmedetomidine as the preferred sedative, at less than 0.6 µg/kg/h with a BIS above 80, on the limited literature available. And the general rule that survives all of it: the degree of suppression tracks the depth of sedation, and agents should be minimised during recording whenever possible.
The effect also varies with the disease, not only the nucleus — the same target behaves differently in Parkinson’s disease and in dystonia — which is a reason to ask the neurophysiologist rather than to generalise.
Monitored anaesthesia care with local anaesthesia and sedation is the most widely used technique, because it permits the monitoring. Local anaesthetic — bupivacaine, ropivacaine or lidocaine, with or without adrenaline — is infiltrated or given as a scalp block for the frame pins and the incision, with more added for closure in a long case.
Conscious sedation is the commonest variant, giving comfort during opening and closure, with short-acting agents stopped before mapping. It suits patients with dystonia or essential tremor who cannot lie still but whose pallidal or Vim recordings are needed. It also produces less postoperative nausea and lets the anti-Parkinson drugs be restarted earlier.
No sedation at all is chosen in some centres, so that nothing influences the recording or the stimulation testing, with analgesia and sedation given only once testing is complete.
General anaesthesia is used for patients who cannot tolerate being awake: paediatric patients, extreme anxiety, developmental delay, severe off-medication symptoms, behavioural disorders. Some centres do every insertion asleep under intraoperative MR guidance, with or without recording, and those centres often do no stimulation testing at all.
None of these has been proven superior. That sentence is the source’s, and it is worth repeating in the anaesthetic room when the technique becomes an argument.
Macrostimulation is the other localising technique and it needs a cooperative awake patient: stimulate the target, see whether the symptoms improve, and watch for side effects from stimulating adjacent structures — paraesthesia, abnormal motor activity — so the electrode can be repositioned. Sedation, if used, must be short-acting and stopped well before testing. General anaesthesia interferes with it.
And one confounder worth naming, because it can mislead the whole team: passing and implanting the electrode produces a microlesion effect around the fresh electrode that can improve the patient’s symptoms with no stimulation at all. Improvement is therefore not by itself proof that the electrode is in the right place.
The stereotactic frame is a navigational tool, and its accuracy is the whole point of the operation. But it introduces problems of its own — MRI distortion related to the frame, discomfort for an awake patient, and the airway consequences that are today’s pitfall.
Frameless techniques achieve similar accuracy to frame-based ones. Robot-assisted placement has been developed to improve accuracy further. Intraoperative MRI is increasingly used for localisation and for confirming placement, with or without a frame, recording or stimulation. If imaging is done outside theatre, the patient must be transported there and back, and may be imaged again afterwards to confirm the electrodes.
Standard monitors — ECG, non-invasive pressure, saturation, end-tidal CO₂ — which can be technically difficult in severe movement disorders and spasticity. An arterial line is considered for severe comorbidity or labile pressure. Oxygen by nasal prongs or mask with a port for end-tidal CO₂ and respiratory rate.
Omit the urinary catheter for the awake patient’s comfort — then watch fluid administration carefully to avoid bladder distension, while still avoiding hypovolaemia.
Depth-of-anaesthesia monitoring is of uncertain usefulness, though bispectral index may help with arousal time, total propofol consumption and cardiopulmonary stability.
Positioning is an airway intervention. Flexion at the lower cervical spine with extension at the atlanto-occipital junction keeps the airway patent and makes it possible to secure in an emergency. Legs flexed and supported under the knees for stability when the patient is sat up. Continuous positive airway pressure may be needed for sleep apnoea. Clear plastic drapes so you can keep verbal and eye contact — and frequent communication, for reassurance and motivation, which is a clinical intervention in a case this long.
Intraprocedural complications occur in 5 to 16% of patients.
| Complication | Rate | Notes |
|---|---|---|
| Intracranial haemorrhage | 0.6–3.3% | Hypertension the only consistent associated factor; target systolic <140 mm Hg |
| Seizure | 0.8–4.5% | Mostly at macrostimulation; usually self-limited and focal |
| Airway and respiratory | 1.6–2.2% | Oversedation, intracranial event, or mechanical |
| Cardiac | 0.4–0.6% | Low overall |
Haemorrhage in an awake patient is suspected from a sudden change in mental status, a new focal deficit, or a sudden onset of hypertension, and confirmed by CT or MRI. Sudden loss of consciousness needs the airway secured, the pressure controlled, and transfer for imaging. The hypertension that raises the risk is often anxiety or poor preoperative control — which sets up a genuine bind, since the obvious treatment for anxiety is sedation, and sedation is the thing that degrades the recording. There is no clean answer to that; the source’s answer is to control the pressure directly and keep the sedation light.
Seizures are mostly at macrostimulation, usually focal and self-limited. A generalised tonic-clonic seizure is treated with propofol 20 mg, and if it is controlled and the patient recovers and is stable, the procedure may continue.
Venous air embolism risk is raised by the sitting or semi-sitting position — the same physiology as Neuro 2 day 09, in a room where nobody expects it. Inform the surgeon and have the field flooded.
Postoperative behavioural and cognitive problems are common, and a restless or violent patient may need urgent care. Look for a cause first — medication, or haemorrhage. Then choose the drug carefully: selective dopamine blockers (clozapine, quetiapine) may be given safely; non-selective blockers (olanzapine, haloperidol) should be avoided.
Internalisation of the electrodes and implantation of the pulse generator is usually done under general anaesthesia. The generator goes in the pectoral area like a pacemaker, with the leads tunnelled up to meet the cranial electrodes.
And a therapy fact that outlives the operation: a patient with Parkinson’s disease whose stimulation stops can develop a potentially life-threatening DBS-withdrawal syndrome, and a patient with generalised dystonia can develop status dystonicus. Urgent intervention to maintain or restore stimulation may be required. If you meet one of these patients later, for any operation at all, the device is not an incidental finding.
The pharmacology has a mechanism, so it can be reasoned from rather than memorised: GABAergic drugs silence GABAergically-driven neurons, which is why the pallidum is harder to record than the subthalamus and why dexmedetomidine won.
The technique has no winner, and saying so is more useful than defending a preference.
And the safety problem is entirely one of preparation. You take away the medication that keeps this patient’s airway working, then fix their head to a table. Both of those are correct. Doing them without arranging the escape route first is not.
No randomized trial addresses this.
The mechanism is established and it does the explanatory work. GABAergic anaesthetic agents affect both background and neuronal spike activity, and the effect is not homogeneous across the brain. Pallidal neurons receive predominantly GABAergic input from the putamen and external pallidum and are therefore more suppressed by propofol than subthalamic neurons. Non-GABAergic drugs — dexmedetomidine, ketamine — and short-acting opioids have minimal effect. That chain of reasoning is why the target determines the technique.
The technique itself is unsettled and the source says so. Monitored anaesthesia care with local anaesthesia and sedation is the most widely used approach because it permits neurophysiological monitoring, general anaesthesia is used where a patient cannot tolerate being awake, and some centres perform every insertion asleep under intraoperative MR guidance. None of the anaesthetic techniques has been proven superior to the others.
The complication rates are observational. Intraprocedural complications in 5 to 16%, intracranial haemorrhage 0.6 to 3.3%, seizures 0.8 to 4.5%, airway and respiratory complications 1.6 to 2.2%, cardiac complications 0.4 to 0.6%. They calibrate vigilance; they do not compare techniques.
One consistent risk factor has emerged from that observational work, and it is the only modifiable one: hypertension is the only factor consistently associated with the occurrence of a haematoma, which is why a systolic target below 140 mm Hg appears in the complications table.
The indication rests on randomized evidence even though the anaesthetic does not. Deep brain stimulation was compared against best medical management in a randomized trial in Parkinson's disease, and that trial is why these patients are on your list — a useful separation to keep, because it is the same shape as Neuro 2 day 08: a well-proven operation whose internal conduct is largely consensus.
Consider what has been done to this patient before the first burr hole.
Their anti-Parkinson medication has been deliberately withheld, because the surgeon needs to see the symptoms in order to test whether stimulation abolishes them. So the tremor, rigidity and bradykinesia are all worse than the person's usual state — and so is everything those symptoms are made of. The source lists what that means for you: bulbar dysfunction, oesophageal dysfunction and gastroparesis raising aspiration risk, restrictive lung disease, and in dystonia an increased risk of laryngospasm, on top of temporomandibular joint and cervical spine rigidity.
Then their head is fixed in a stereotactic frame, and the frame is fixed to the operating table.
Conventional laryngoscopy may now be awkward or impossible. The frame often restricts neck movement, and it may cover some or all of the mouth and nose. You have made the patient's airway worse on purpose and then removed your usual means of rescuing it — and if you have not thought about that in advance, you will be thinking about it while somebody is obstructing.
There are three routes into this emergency and it is worth knowing all of them, because they do not all announce themselves the same way.
Oversedation. Respiratory depression from sedative given to keep an anxious patient still. Airway and respiratory complications occur in 1.6 to 2.2% of these procedures.
An intracranial event. A seizure or a haemorrhage with a falling conscious level. Here the airway problem is a symptom, and treating it without recognising the cause loses time on the thing that will actually injure the patient.
Mechanical obstruction with no drug involved at all. A restless awake patient moves their body while the head stays fixed to the bed. Nothing has been given, nothing has bled, and the airway obstructs because the neck has been levered into a position the patient cannot correct.
The response is specific and slightly counter-intuitive.
Try to secure the airway without removing the frame. A laryngeal mask, or videolaryngoscopy with the frame in situ, is the recommended first move — and the reason is not only speed. If the airway is secured with the frame undisturbed, the surgery can continue. Take the frame off and the stereotactic coordinates go with it, and the patient has had a craniotomy for nothing.
But have the release equipment in the room. Equipment to release the frame from the table and to remove it from the patient must be immediately available. Planning to avoid using it is not the same as not having it.
Two more details close the loop.
If the case is being done asleep, intubate before the frame goes on. When general anaesthesia is used, induction and intubation may take place before the headframe is placed — which converts the whole problem into a non-problem. The sequence is the safeguard.
And the drugs come back afterwards. Respiratory insufficiency can occur postoperatively simply from the continued absence of anti-Parkinson medication. Conscious sedation is preferred partly because it lets those drugs be restarted earlier. In a patient who already has a stimulator, failure to maintain or restore stimulation can cause a potentially life-threatening DBS-withdrawal syndrome, and in generalised dystonia, status dystonicus — so restoring the therapy is itself an urgent intervention, not an administrative step.
Full text · Preston goes through LibKey to whatever Preston holds — the PDF where there is one — and works on a phone. The doi link always reaches the publisher's record, whether or not we hold a subscription. If you are reading on a computer, LibKey Nomad adds the same one-click PDF button to PubMed and publisher pages everywhere else; it is a desktop browser extension, so it does nothing on a phone or tablet.
A patient with dystonia is scheduled for pallidal deep brain stimulation with microelectrode recording. The surgeon asks whether the case can be done under general anaesthesia with propofol. What is the objection?
Midway through an awake deep brain stimulation insertion, a restless patient shifts their body while the head remains fixed in the frame, and the airway obstructs. What should you do first?
Which statement about intracranial haemorrhage during deep brain stimulation insertion is correct?
Ascending transtentorial herniation pushes the pons, vermis and adjacent cerebellum up through the incisura, effacing the cisterns symmetrically and pressing the upper pons forward against the clivus. Aqueductal compression frequently produces acute hydrocephalus on top of the mass effect that caused it.
Displacement of the septum pellucidum from the midline is predictive of prognosis in subfalcine herniation, and a correlation between midline shift and level of consciousness has been shown. Vertical descent is the one that behaves less predictably. So the imaging tells you reliably what is happening and less reliably how bad it is.