An external ventricular drain is two instruments that cannot be used at the same time. Open, it is a pressure prescription written in centimetres of water. Closed, it is a monitor. Most of the errors come from forgetting which one it currently is.
The height of the drip chamber above the foramen of Monro is the pressure the patient is allowed to keep — raise the chamber and you have raised the permitted intracranial pressure. And while the drain is open the transducer is reading that height, not the patient, so a number charted with the drain open is not an intracranial pressure. The system produces about 0.35 to 0.40 mL of CSF a minute whatever you do with the stopcock, which is why the decision is never whether the fluid keeps arriving.
Cerebrospinal fluid keeps arriving. Measured non-invasively at the cerebral aqueduct by phase-contrast MRI, production runs at 0.35 to 0.40 mL/min — about 500 to 600 mL a day. Roughly 0.25% of the total adult volume is replaced every minute, so the whole volume turns over every 5 to 7 hours, about four times a day. The choroid plexus supplies the majority, around 50 to 80%, and the rate is under circadian control and falls with age.
None of that pauses because a stopcock is closed. It is the reason a drain is a continuous decision rather than a procedure that was performed once.
An external ventricular drain does two jobs that share one lumen, and it can only do one at a time.
Open, it is a therapy. Fluid leaves whenever the intracranial pressure exceeds the height of the drip chamber above the zero reference. That height is therefore a pressure prescription — it is the pressure the patient is allowed to keep. Raising the chamber raises the permitted pressure; lowering it drains more and risks over-drainage.
Closed, it is a monitor. Now the fluid column is continuous with the patient, the waveform belongs to them, and the mean is a real intracranial pressure.
The trouble is that both states produce numbers, and only one of them produces numbers about the patient.
Everything the device tells you is relative to a point you choose. The transducer is zeroed at the level of the ventricles: the tragus of the ear in a supine patient, and the midsagittal line between the eyebrows in the lateral position.
Two things follow. The reference moves when the patient moves, so sitting someone up or turning them changes both the recorded pressure and the amount of fluid leaving their head, without anyone touching the equipment. And a qualitative assessment of the ICP tracing is what confirms a reading is reliable — a damped or absent waveform means the catheter is not communicating with the ventricle, and the number on the screen is an artefact carrying a decimal point.
If a lumbar drain is used instead to estimate pressure, the patency of the CSF pathways should be established first with the Queckenstedt manoeuvre — lumbar CSF pressure rising when the jugular veins are compressed. Without a patent pathway you are measuring a disconnected compartment.
The intraventricular catheter has historically been called the gold standard for ICP monitoring, and intraparenchymal microtransducers give equivalent pressure measurement with lower risk of haematoma and infection. So the argument for a ventricular catheter is not accuracy. It is that a ventricular catheter can also drain, and drainage is a treatment the parenchymal device cannot offer.
Their weaknesses are different in kind, which is worth knowing when you are told one is better. The microtransducer suffers zero drift over several days and cannot be recalibrated in vivo. The ventricular catheter can be re-zeroed at the bedside and carries the risks that come with entering the ventricle.
The Neurocritical Care Society published an evidence-based consensus statement on the insertion and management of these devices, and the reason such a document exists is instructive: the everyday questions — continuous or intermittent drainage, when to clamp, how long to leave the catheter, how to prevent infection — had accumulated a great deal of local custom and rather less evidence. Knowing that the statement exists, and that it is a consensus statement, is more useful than memorizing any single recommendation from it.
Two practical points survive independently of the debate.
Clamp before you move the patient. A transfer changes the geometry between the head and the chamber, which is the entire reference frame. An open drain during a move can siphon; a drain left clamped afterwards stops working quietly.
Continue monitoring after a large resection. Following removal of large masses with extensive peritumoral oedema, intracranial hypertension requiring prompt intervention develops within hours in up to 20% of patients, from swelling or haematoma. Postoperative monitoring also helps clarify the cause of a delayed emergence, which is otherwise a diagnosis made by exclusion in a patient who cannot contribute to it.
It is worth carrying day 03’s finding into this one. The only randomized trial of ICP monitoring in traumatic brain injury found no difference in outcome between a protocol built around a monitor and one built around serial imaging and examination. That result does not argue against drainage, which is a treatment with a mechanical rationale rather than a threshold. It argues against the belief that having the number is the same as having done something with it.
No randomized trial addresses this.
The physiology is settled and quantified. Cerebrospinal fluid production measured non-invasively at the cerebral aqueduct by phase-contrast MRI runs at 0.35 to 0.40 mL/min, or roughly 500 to 600 mL a day. About 0.25% of the total adult CSF volume is replaced each minute, so the turnover time for the whole volume is 5 to 7 hours — around four times a day. The choroid plexus produces the major portion, roughly 50 to 80%, with the remainder from non-choroidal sources. Production is under circadian control and falls with age.
The management is consensus. The Neurocritical Care Society published an evidence-based consensus statement on the insertion and management of external ventricular drains precisely because the practice questions — continuous versus intermittent drainage, when to clamp, how to prevent infection, when to remove — had accumulated more custom than evidence. It is the right document to know exists, and it is a consensus statement rather than a set of trial results.
Two facts about the device are not in dispute. The transducer must be zeroed at the level of the ventricles — the tragus of the ear in a supine patient, the midsagittal line between the eyebrows in the lateral position. And a qualitative look at the ICP tracing is necessary to confirm the reading is real, because a damped or absent waveform means the number is not measuring anything.
Strength, in the guideline's own words The Neurocritical Care Society document is styled an "evidence-based consensus statement," not a guideline, and its recommendations are graded accordingly.
This is the commonest error with the device and it is almost invisible, because nothing about it looks wrong. The ICP column on the chart fills up hourly with plausible numbers, they are low, and everyone is reassured.
While the drain is open, the fluid column is referenced to the drip chamber. The transducer is reading the height at which you hung it. If the chamber sits 15 cm H₂O above the zero point, the displayed pressure will sit somewhere near that regardless of what the brain is doing — because any pressure above it is being relieved, continuously, by the drain. The measurement and the treatment are the same tube, and they cannot both be happening.
Two habits fix it. Chart pressures only from a closed system, and record in the same entry that it was closed. And look at the waveform: with the drain open the trace is damped or absent, which is the tell, and a qualitative assessment of the tracing is the recommended way to confirm a reading is reliable in the first place.
The related trap is the reference point rather than the stopcock. The zero is the level of the ventricles — the tragus supine, the midsagittal line between the eyebrows in the lateral position. Sit a patient up, or turn them, and the reference has moved even though nobody touched the equipment. The same error in the arterial line makes CPP read higher than it is; here it changes both the number and the amount of fluid leaving the head.
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 has an external ventricular drain open to drainage at 15 cm H₂O. The chart records an intracranial pressure of 13 to 15 mm Hg every hour for six hours. What should you conclude?
Which statement about external ventricular drains and cerebrospinal fluid physiology is correct?
A team decides to raise the drip chamber of an external ventricular drain from 10 to 20 cm H₂O. What is the consequence?
CO₂-induced cerebral vasoconstriction wanes over roughly 6 to 10 hours as bicarbonate is exchanged and CSF pH normalizes at the low PaCO₂. The vessels are no longer constricted at 25, so returning the PaCO₂ to 40 raises flow above where it started. The correction, not the state, is the moment that carries risk — and the same logic in reverse applies to the chronically hypercapnic patient.
The source text describing the 50–150 plateau says so itself — it is "only a statistical description of how the general population responds," a value of 50 does not guarantee any particular patient is on the plateau, individual responses vary widely, and the relevance of the idealized curve has been questioned. Drummond's review is that questioning made specific. The practical consequence is not a new number to memorize but the loss of the old one's authority.