Neurapheresis in Neurotechnology: Filtering the Fluid That Bathes the Brain

Most neurotechnology of the last forty years has been electrical. Electrodes record. Stimulators modulate. Implants try to push a misfiring circuit back toward something like normal.

Neurapheresis is a different bet. It treats the nervous system as a fluid system as well as an electrical one. Instead of talking to neurons with current, it circulates and filters cerebrospinal fluid (CSF), the liquid that surrounds the brain and spinal cord. The working analogy is dialysis: not for blood, but for the central nervous system.

That places Neurapheresis in a small but growing part of neurotech that is less about chips and more about plumbing, chemistry, and waste clearance.

The problem Neurapheresis is built to solve

CSF is not inert packing fluid. It cushions the brain, helps regulate intracranial pressure, carries nutrients, and clears metabolic waste. It also becomes a transport medium for things that damage the nervous system:

- blood and haemoglobin after aneurysmal subarachnoid haemorrhage

- organisms and antigen in meningitis

- tumour cells in leptomeningeal metastases

- inflammatory proteins and cellular debris after injury or infection

Blood has dialysis. Plasma has apheresis. CSF, until recently, has had drains, shunts, and lumbar punctures: ways to remove volume or sample fluid, not closed-loop ways to clean it and put it back.

That gap is the neurotech argument. If disease leaves a signature in CSF, or if the fluid itself becomes toxic, then a device that can move, filter, and return CSF is a platform, not a one-off procedure.

What Neurapheresis is

Neurapheresis™ is a catheter-based, extracorporeal CSF management system. In published descriptions it:

1. withdraws CSF from the lumbar cistern through one lumen of a dual-lumen spinal catheter

2. passes that fluid through a filtration unit outside the body

3. returns filtered, autologous CSF through the second lumen, typically into the thoracic subarachnoid space

4. sends removed material to a waste path rather than back into the patient

The system is designed to run as a closed loop. The aim is to process CSF continuously for hours, keep volume more stable than simple drainage, and avoid replacing large amounts of native fluid with artificial substitutes.

The name is literal. “Neur” refers to the nervous system. “Apheresis” means separating a component from a mixture. In patents and papers, Neurapheresis is defined as modifying CSF contents: removing cells, pathogens, blood products, inflammatory mediators, or other targets, and in principle circulating or adding agents as well.

The technology originated at Minnetronix Neuro. It is now being advanced by Pharaoh Neuro, a Minneapolis company that in February 2026 announced a $20 million Series A, with Amaza Reitmeier, formerly of Medtronic neuromodulation, as CEO and Duke functional neurosurgeon Nandan Lad as chief medical officer. Lad has described the idea in public as neurodialysis.

Where it sits in the neurotech map

It helps to sort neurotechnology by what it actually touches.

Approach	What it acts on	Typical tools
Recording and decoding	Electrical activity	EEG, ECoG, Utah arrays, BCI threads
Neuromodulation	Circuits and pathways	DBS, SCS, DRG stimulation, focused ultrasound, TMS
Drug and gene delivery	Molecular targets	Intrathecal pumps, viral vectors, focused-ultrasound BBB opening
Structural access	Anatomy	Endoscopes, ports, robots
Fluid and chemical clearance	CSF and its contents	Drains, shunts, Neurapheresis

Neurapheresis belongs in the last row. It does not read intention or stimulate a nucleus. It changes the composition of the extracellular fluid compartment that neurons and glia live in.

That is why comparisons to Neuralink-style implants miss the point. A BCI is an interface. Neurapheresis is closer to an organ-support device for the CSF space, in the same family as dialysis machines, ECMO, or plasma-exchange systems, adapted to a much smaller, more pressure-sensitive compartment.

There is a related research idea sometimes called a “CSF sink”: if pathogenic proteins can be pulled out of CSF, the brain’s interstitial fluid may follow. Implantable CSF-apheresis concepts for amyloid in Alzheimer’s models sit in that same conceptual neighbourhood. Neurapheresis is the acute, extracorporeal, catheter-based version of that logic, built first for critical care rather than chronic outpatient protein clearance.

How the technology works in practice

The access is spinal, not cranial. After the primary neurosurgical problem is dealt with where needed — for example after an aneurysm is secured — a dual-lumen catheter is placed at L3–L4 or L4–L5 and can be advanced into the upper thoracic CSF space over a wire under fluoroscopy.

A pump and filter cassette then create a flow loop. Published engineering and trial descriptions use aspiration and return rates on the order of a couple of millilitres per minute, far higher than a typical lumbar drain. Computational and in-vitro work has argued that this loop changes mixing and clearance along the spinal subarachnoid space more than passive drainage does, with the strongest effect between the aspiration and return ports.

Filters can be chosen for the job. Blood and cell removal after haemorrhage is a different pore-size and membrane problem from yeast removal in cryptococcal meningitis or circulating tumour-cell reduction in leptomeningeal disease. The platform idea is that the catheter and pump stay recognisable while the filter chemistry changes with the target.

Two design choices matter for neurotech readers:

- **Autologous return.** The system tries to give the patient back their own filtered CSF instead of discarding fluid and replacing it. That is a pressure- and biochemistry-preserving move in a compartment where volume errors are dangerous.

- **Closed loop.** A sealed circuit is meant to reduce infection risk and uncontrolled loss compared with open drainage, though any indwelling spinal catheter still carries infection, leak, and placement risk.

The evidence, without inflating it

The most important human data are in aneurysmal subarachnoid haemorrhage.

PILLAR was the first-in-human study: catheter placement after aneurysm treatment, filtration for up to 24 hours, measurable drops in CSF red cells and protein, and imaging signals of less cisternal blood. It was a safety and feasibility study, not a pivotal outcomes trial.

PILLAR-XT extended duration and sample. Published results reported successful catheter insertion in 27 of 29 participants, a median filtration time around 37 hours, and large mean reductions in CSF red blood cells and protein compared with published standard-of-care trajectories. Captured adverse events in that series were described as mild or moderate and resolved. Again, the study was built to test whether the system can be used and whether it clears blood products, not to settle long-term disability or mortality.

Outside haemorrhage, the literature is earlier:

- in-vitro and rabbit work in cryptococcal meningitis showed rapid mechanical reduction of yeast and antigen

- in-vitro work in leptomeningeal metastases showed log-scale reduction of tumour cells and better mixing of intrathecal drug in a model spine

Those are reasons to keep developing the platform. They are not reasons to treat Neurapheresis as a proven therapy for infection or cancer.

Regulatory status is straightforward. The system has been used under investigational-device conditions. Pharaoh Neuro has said proceeds from its Series A will support development and FDA clearance. It is not a cleared consumer or pain-clinic product.

Why this is neurotechnology, not just a better drain

A lumbar drain already moves CSF. Neurapheresis adds three neurotech ingredients.

First, **active control.** Flow rate, filtration, and return become programmable rather than gravity-dependent. That is the difference between a tap and a circuit.

Second, **selectivity.** The filter is a design surface. In principle one can target cells, haemoglobin, pathogens, or selected proteins. Patents around the platform list a wide set of possible future conditions — meningitis, encephalitis, selected neurodegenerative diseases, spinal cord injury, traumatic brain injury, vasospasm — as the kind of list device companies write when they think they have a platform. Most of that list is still aspiration, not evidence.

Third, **data.** A closed CSF loop can sample what it removes. That makes the device a potential diagnostic as well as a treatment: a way to watch clearance curves, inflammatory markers, or pathogen load in near real time. For neurotech, a therapy that also generates a high-resolution chemical time series is more interesting than a silent drain.

Put differently, neuromodulation edits signals. Neurapheresis edits the medium the signals live in.

Limits and risks

The honest constraints are as important as the pitch.

- The human evidence base is still small and concentrated in one emergency indication.

- Catheter placement is an invasive spinal procedure in very sick patients.

- Risks include infection, bleed, nerve injury, CSF leak, malposition, over- or under-drainage, and device failure.

- CSF volume is small and pressure is tightly regulated. Flow algorithms matter; this is not a forgiving circuit.

- Clearing a solute from lumbar CSF does not automatically clear the same solute from brain parenchyma. The spinal and cranial compartments do not mix like a stirred beaker.

- Broader uses — neurodegeneration, chronic inflammation, outpatient “CSF cleaning” — remain conceptual until someone runs the right trials.

There is also a language problem. Because “neurapheresis” sounds like “filtering nerves,” it is easy to smuggle the term into pain-clinic marketing. That is not what the device does. Blood plasma exchange for some autoimmune neuropathies is a different procedure in a different compartment.

What to watch next

For people who follow neurotech rather than only neurosurgery, the next markers are practical:

- whether Pharaoh can take the system from investigational use to a defined FDA pathway

- whether aSAH studies move from clearance curves to functional outcomes

- whether filter sets for infection or leptomeningeal disease enter prospective human studies

- whether anyone pairs CSF filtration with drugs, antibodies, or diagnostics in the same loop

- whether chronic, lower-intensity CSF apheresis stays science fiction or becomes a separate implant class

The deeper question is architectural. Neurotechnology has spent a generation on electricity. Waste, inflammation, blood products, and protein load in CSF are a different control layer. Neurapheresis is one of the first serious industrial attempts to build a machine for that layer.

It will succeed or fail on clinical outcomes, not on the elegance of the analogy. But the analogy is the right one: if the brain has electricity, it also has plumbing. This is a device for the plumbing.