26.3 Complications of EVD and Lumbar Drain (04.G, I)
Key Takeaways
- EVD infection and ventriculitis become more common the longer the catheter stays in; a prospective UK and Ireland cohort found a 9.3% infection risk, with duration of 8 days or more roughly doubling the hazard, and frequent CSF sampling adding further risk.
- Overdrainage produces subdural hygromas or hematomas and, when a tight posterior fossa remains, upward (reverse) herniation. Underdrainage and occlusion present as a dry bag, a rising examination, and a climbing ICP.
- Zero the transducer at the tragus or external auditory meatus (foramen of Monro). An incorrectly zeroed drain fabricates ICP. Never clamp an EVD — or clamp an EVD and a lumbar drain together — and leave the patient without ICP data or frequent examinations.
- A lumbar drain is contraindicated in obstructive hydrocephalus or a crowding posterior-fossa mass. Nerve-root pain and retained catheter fragments are device-specific harms.
- The same perioperative watch list includes sodium shifts (SIADH, cerebral salt wasting, diabetes insipidus) and positioning injuries (ulnar and brachial-plexus palsies, peroneal palsy, sitting-position venous air embolism, rare ischemic optic neuropathy after prone surgery).
An external ventricular drain (EVD) is a straw in a pressurized box. Independent OpenExamPrep teaching for outline items 04.G and 04.I is the ways that straw injures: tract hemorrhage, overdrainage, underdrainage, occlusion, infection, disconnection, and wrong zero. A lumbar drain (LD) is not a smaller EVD. It can herniate a patient with obstructive hydrocephalus, sting a nerve root, and snap off in the canal. Height, weaning, and when an LD is anatomically legal were taught with hydrocephalus in chapter 23.3; this section is the complication list the perioperative outline names, plus the sodium and positioning injuries that arrive on the same ICU day.
Tract hemorrhage and the bloody start
Passing a ventricular catheter through brain can bleed along the tract or into the ventricle. Risk rises with multiple passes, coagulopathy, antiplatelet drugs, very small ventricles, and a hurried bedside stick without image guidance. Most tract bleeds are small and watched. Expanding tract or intraventricular hemorrhage with a declining examination is a scan-and-correct-coagulation problem; a second pass through a fresh hematoma is how you make a third pass. Do not “just lower the drain to 0” to empty a new tract hematoma — you will overdrain the clean ventricle and still not evacuate a clot.
Overdrainage, underdrainage, and occlusion
Overdrainage is CSF leaving faster than production (about 20 mL/h in an adult) or faster than the brain can stay supported. Ventricles slit, the cortical mantle pulls off the inner table, and subdural hygromas or hematomas appear — the same overdrainage physiology as a too-low shunt valve. Headache that is worse sitting, a sinking examination after you dropped the burette, and new extra-axial collections are the clues. In a tight posterior fossa (cerebellar hematoma, infarct, or tumor still in place), unloading the supratentorial ventricles too fast lets the cerebellum move up through the tentorial notch: upward (reverse) herniation. Pupils go midposition, the patient drops, and the “helpful” EVD at 0 cm H2O was the mechanism. Raise the drain, give volume if they are dry, and decompress the posterior fossa — do not drop the drain further.
Underdrainage is the opposite: the bag is dry, ICP (if you are measuring it) is high, the examination falls, and the CT ventricles are still large. Causes: the stopcock is off, the drip chamber is above the pressure, the catheter is against the choroid, or the lumen is clotted. A dry bag is never proof that hydrocephalus resolved.
Occlusion is blood, tissue, or a kink. Output falls over minutes to hours. Flush only with a protocolled small sterile volume; this is not a kitchen-sink irrigation. Persistent occlusion means replace the catheter. Intraventricular thrombolysis protocols are a selected-IVH research or center pathway (chapter 20), not a routine unclog maneuver on a tumor EVD.
| Problem | What you see | First correction |
|---|---|---|
| Overdrainage | Low output needed, subdural collections, sitting headache, upward herniation if the posterior fossa is tight | Raise the chamber; stop suction; decompress posterior fossa if needed |
| Underdrainage | Dry bag, high ICP, large ventricles, decline | Open the stopcock, lower a too-high chamber, unkink |
| Occlusion | Sudden drop in output, waveform dies, flushing fails | Sterile protocol flush or replace |
| Wrong height | “ICP 25” with a smiling patient, or a dying patient with “ICP 6” | Re-zero at the tragus / external auditory meatus |
1 mm Hg ≈ 1.36 cm H2O. A drain open at 15 cm H2O opens near 11 mm Hg. Wrong zero is a fabricated number. The transducer follows the foramen of Monro when you sit the patient up; leaving it taped to the IV pole at bed height while the head is at 45° is how “ICP” becomes fiction.
Infection and ventriculitis: duration is the risk you can name
A prospective multicentre UK and Ireland study (495 catheters, JNNP 2018) found an EVD-related infection risk of 9.3%. Duration ≥8 days independently increased risk (hazard ratio 2.47). Daily or every-other-day sampling increased risk about five-fold compared with sampling once or twice a week. Catheter coating (antibiotic-impregnated versus silver versus plain) did not show a significant infection difference in that cohort — do not hang a stem on “the silver catheter cannot infect.” Older surveillance series put device-associated meningitis near 6 infections per 1,000 EVD-days and higher per lumbar-drain-day; the exact integer varies with definition. The exam-stable facts: risk rises with days, breaks in the closed system, CSF leak at the tunnel, and needless sampling. Routine prophylactic catheter exchange on day 5 is not a proven infection cure and costs another tract hemorrhage.
When infection is suspected: sample from the drain (Gram stain, culture, cell count, glucose, protein), remove the infected catheter, replace at a new site if diversion is still required, and start vancomycin plus cefepime, ceftazidime, or meropenem (IDSA 2017). Interpretation is messy in bloody CSF; a falling glucose and a rising neutrophil count still matter. Chapter 23.4 is the drug table; this section is why day 10 is more dangerous than day 2.
Disconnection, clamps, and the sentence you should not have to read twice
Accidental disconnection opens a sterile ventricle to the bed sheet and can dump CSF on the floor (overdrainage plus infection). Clamp the proximal catheter with a sterile technique, call neurosurgery, and do not “just reconnect the dirty tubing.”
Never clamp an EVD and fail to watch ICP. Clamping is a wean or a transport step, not a convenience so the bag stops filling. If you clamp, you need either a transduced ICP or q15–q30-minute examinations and a low threshold to unclamp. If the patient has both an EVD and a lumbar drain, clamping both without a monitor is how you miss a pressure spike. Transport with a clamped EVD and no waveform is a classic night-shift disaster.
Society for Neuroscience in Anesthesiology and Critical Care perioperative drain guidance (Lele and colleagues, 2017) is the document many operating rooms use for how to handle drains during induction and positioning: know whether the drain is open or transduced, do not let it siphon to the floor when you sit the table, and do not assume the number is real if the zero wandered.
Lumbar drain: herniation, roots, retained catheter
A lumbar drain is contraindicated when CSF pathways do not communicate: obstructive hydrocephalus, a crowded posterior fossa, uncal or tonsillar herniation patterns, or a large supratentorial mass with a pressure gradient to the spine. Removing lumbar CSF increases that gradient. Tonsils descend, or the cerebellum moves up if you have also emptied the ventricles — both are herniation. If you need diversion in those anatomies, the tool is a carefully managed EVD plus surgical decompression, not a lumbar catheter.
Nerve-root irritation is radicular pain, dysesthesia, or a dropped foot after insertion or after the catheter migrates. Pull back or remove; do not keep draining through a screaming L5 root. Retained catheter happens when the tube shears on a needle or on the fascia during a fight. Image the fragment (CT) and tell neurosurgery; a small asymptomatic fragment is sometimes observed, a symptomatic or infected fragment is not a souvenir.
Overdrainage from a lumbar drain is common (some series report overdrainage symptoms in a large fraction of patients): low-pressure headache, brain sag, cranial-nerve palsy, and pneumocephalus. Target volumes are often in the 5–15 mL/h range, not an open stopcock on the floor.
Other perioperative issues folded here: sodium and positioning
Sodium shifts travel with the same operations that get drains. After pituitary or hypothalamic work, diabetes insipidus is high urine output and rising sodium — replace free water and give desmopressin when the diagnosis is clear (chapter 11.1). After SAH or hemispheric injury, SIADH (euvolemic hyponatremia, inappropriately concentrated urine) and cerebral salt wasting (hypovolemic hyponatremia, high urine sodium) are the pair you must not treat with the same bag (chapter 11.2). A falling sodium is also a seizure risk that mimics “early postoperative epilepsy.” Check the chemistry before you add a third ASD.
Positioning injuries are the complications that are not on the CT. Sitting or semi-sitting posterior-fossa surgery: venous air embolism, tongue and face swelling after a long flexion, and rare mid-cervical cord ischemia from extreme neck flexion. Supine with the arm abducted: brachial plexus and ulnar palsies. Lithotomy or poorly padded fibular head: common peroneal palsy. Prone spine cases: ischemic optic neuropathy (painless visual loss, often after long blood-loss cases with hypotension). Mayfield pins can, rarely, cause a skull fracture or epidural hematoma. Document a focal peripheral deficit as a positioning or pin injury after you have ruled out an intracranial disaster — not instead of the CT.
Exam traps
Celebrating a dry EVD bag. Zeroing the transducer to the IV pole. Clamping for a trip to CT and forgetting to look at the patient. Lumbar-draining a closed fourth ventricle. Daily CSF “surveillance cultures” on a well patient. Flushing an occluded EVD with 20 mL because the resident is strong. Blaming every hyponatremia on the drain height. Missing a peroneal palsy because you never looked at the feet.
In practice
EVD day 9, fever, cloudy CSF: sample, new catheter, vancomycin plus cefepime — duration did this. Posterior-fossa infarct, EVD open at 0 cm H2O, sudden coma and midposition pupils: upward herniation from overdrainage, raise the drain and decompress. Clamp for MRI, no ICP cable, 40 minutes in the hallway: unclamp or transduce — that clamp was not a break. Lumbar drain for “hydrocephalus” with a cerebellar burst and an invisible fourth ventricle: pull the lumbar plan, place an EVD. Sitting craniotomy, then a wrist drop: CT first, then call it ulnar palsy if the scan is clean. Sodium 118 mmol/L and a first “postop seizure”: treat the sodium and the seizure together.
An EVD is clamped for a trip to CT. There is no transduced waveform, and no one is assigned to watch the examination. Which statement is correct?
A patient with a large cerebellar infarct has an obliterated fourth ventricle. The team proposes a lumbar drain “to rest the posterior fossa.” What is the correct complication warning?
Which pair about EVD infection timing and transducer zero is accurate?