Hydrocephalus: the fluid pressure on the brain, the shunt that drains it, and the NPH that hides as dementia

Last updated September 3, 2026.

Hydrocephalus is a build-up of cerebrospinal fluid inside the brain, raising pressure where there should be free flow. It has two famous faces: the baby whose head grows too fast because the fluid cannot drain, and the older adult whose walking, bladder, and memory slip together in a pattern easily misread as dementia or Parkinson's. Both deserve to be known, because both are treatable. The treatment is drainage: a shunt, a thin tube that carries the fluid from the brain to the abdomen, where the body absorbs it, or in some babies a camera procedure that opens a new channel inside the brain. Shunts are lifelong companions for many: they work, they need watching, and they occasionally block or infect, which is why every family and every patient learns the blocked-shunt symptoms by heart. The older-adult form, normal pressure hydrocephalus, is the great missed diagnosis of its age group: the triad of a slowed walk, urinary urgency, and memory trouble, when recognized, can be substantially reversed by a shunt, which makes it one of the few treatable causes of a dementia-like picture. The worth-knowing part: outcome depends on timing, and timing depends on recognition, so the pattern is worth more than the word.

What does it look like?

In babies: a head growing too fast across the centile lines, a bulging soft spot, downward-turned eyes, vomiting, sleepiness, and irritability. In the older adult: the triad, a walk that has become slow, wide, and glued to the floor, a bladder that has become urgent, and a memory and thinking that have dulled, usually over months. The walking change typically arrives first and is the most reversible when treated early.

Find answers to hydrocephalus

Start a free AI doctor consult →

Why does it happen?

The cerebrospinal fluid is made continuously and must drain and absorb continuously; hydrocephalus is any break in that balance. In babies: a channel too narrow from development, bleeding after premature birth, or spina bifida. In adults: bleeding, infection, tumors, or head injury, and in the older-adult form, an absorption that fails with age for reasons not fully understood. It is nobody's fault in any of its forms.

How is it treated?

When is it urgent?

A blocked or infected shunt is the emergency: headache with vomiting, unusual sleepiness or confusion, fever in someone with a shunt, or the original symptoms returning are each a same-day trip to the emergency department. Pymander's escalation routing is built and tested specifically for this class of decision; see the safety architecture working paper.

What a Pymander AI doctor consult looks like

Do you or does someone you care for have hydrocephalus or a shunt?
My dad is 74 and over the past year his walking has become a shuffle, he has started having bladder accidents, and his memory has gone downhill. We had accepted it as dementia starting. Last week a neurologist said it might be normal pressure hydrocephalus and they want to do a lumbar puncture test. I looked it up and it says a shunt could reverse it. I am terrified to hope. How can something that looks exactly like dementia be fixable, and if it is fixable, why did nobody see it a year ago?
Thanks for giving me all this information. Based on what you've said, let me answer the two questions honestly, because they are the right two. How can it look exactly like dementia and be fixable: because normal pressure hydrocephalus is, by design of bad luck, a perfect mimic. The fluid builds up slowly, the pressure readings can even be normal, and the three signs it produces, the shuffling walk, the bladder urgency, the dulled memory, are the three things every family and many doctors file under getting old or getting dementia. The differences are in the pattern: the walking usually changes first and most, the memory problem has a slowed, not-lost quality rather than the erasures of Alzheimer's, and the three arriving together over a year is the classic shape. It is one of the very few dementia-like conditions with a treatment that can reverse it, which is why the neurologist is testing rather than guessing. Why nobody saw it a year ago: because it is rare next to dementia, because its signs overlap so completely, and because the scan changes are easy to read as ordinary shrinkage until someone puts the walking, the bladder, and the memory into one picture. That miss is common enough to have a name in the literature; it is not negligence, and it is not your miss. On the hope you are terrified of: you are allowed to hope, with the structure the test gives you. The lumbar puncture test, walking measured before and after a fluid drain, is exactly the instrument for calibrating hope: a clear improvement after the drain predicts a good shunt response, and the walking, the first thing that went, is often the first thing back. Some families describe the days after a successful shunt as getting their person back. Not everyone responds, and the test exists to find out which side your dad is on before any operation. Whatever the result, the year was not wasted: it brought you to the right office. Let the test speak, and let yourself hope in proportion to it. Here's his care note to share with his care team.
Care note
NPH 74M son - mimic explained, the miss normalized, hope calibrated to the tap test, walking first back
Son of seventy-four-year-old man: a year of shuffling walk, bladder accidents, and memory decline, accepted as dementia, neurologist now suspects normal pressure hydrocephalus and wants a lumbar puncture test, son has read a shunt could reverse it and is terrified to hope, asks how a dementia mimic can be fixable and why nobody saw it: the diagnostic-crossroads consult. Plan: the mimicry explained (rare beside dementia, perfect overlap, scan changes read as shrinkage), the year's miss normalized as the condition's known pattern rather than negligence, hope explicitly permitted and calibrated to the tap test (walking measured before and after, walking first to return), the honesty that not everyone responds, and the year reframed as the route to the right office.
View care note →

Illustrative example, not a real member's messages.

Common questions

How can this look like dementia but be treatable?

Because normal pressure hydrocephalus is a near-perfect mimic, and that is its known danger. The slowly building fluid produces exactly the three changes families file under aging or dementia: the shuffling walk, glued to the floor,, the bladder urgency, and the dulled memory. The differences live in the pattern: the walking usually changes first and most, the memory problem is more slowed than lost, and the three arriving together over months is the classic shape. It is one of the few dementia-like conditions a treatment can reverse, which is why the neurologist tests rather than guesses: the lumbar puncture drain, with walking measured before and after, is the instrument that separates the mimic from the real thing.

Why did nobody spot it a year ago?

Because it is rare next to dementia, because its signs overlap almost completely, and because its scan changes are easy to read as ordinary age-related shrinkage until someone assembles the walking, the bladder, and the memory into one picture. The miss is common enough that the condition has a reputation in the medical literature for being found late, sometimes years late, and the people who miss it include good doctors. It is not negligence, and it is not the family's miss either: you brought him to the office that assembled the picture, and the year was the route there, not time wasted. The question that matters now is the test result, not the calendar.

Should I let myself hope about the shunt?

Yes, with the calibration the test gives you, and that is precisely what the test is for. The lumbar puncture drain, with his walking measured before and after, predicts the shunt response: a clear improvement after the drain is the strongest sign a shunt will help, and the walking, the first thing that went, is often the first thing back. Families of responders sometimes describe the weeks after a successful shunt as getting their person back. The honesty on the other side: not everyone responds, the memory improves less reliably than the walking, and the test exists to find out which side he is on before anyone operates. Hope in proportion to the test, and let the test speak.

What does the lumbar puncture test actually involve?

A needle into the fluid space in the lower back, under local anesthetic, draining a measured amount of the cerebrospinal fluid, with his walking and sometimes his thinking timed before the drain and again after, over the following hours to days. The team is looking for a visible change, usually in the walking first. It is a day procedure, the headache afterward is the commonest side effect and passes, and the fluid refills itself within hours. Some centers use a longer version, a small drain left in for a few days, when the single drain leaves the answer unclear. It is the least invasive way of asking the shunt question before anyone commits to surgery.

What is life with a shunt actually like?

For most people, surprisingly ordinary, with one discipline. The shunt sits under the skin, a thin tube from the brain to the abdomen with a valve, and once healed it is mostly invisible to daily life: no activity restrictions for most, occasional checks with the surgical team, and a long working life for the device. The one discipline is knowing the blocked-shunt symptoms by heart, headache with vomiting, unusual sleepiness, fever, or the old symptoms returning, and treating them as a same-day emergency every time, because a blocked shunt lets the pressure climb. Families make it second nature within months, and then the shunt fades into the background where it belongs.

If the test does not show improvement, what then?

Then the test has done its job: it has saved him an operation that would not have helped, and the diagnosis returns to the dementia pathway with one important difference, it has been properly checked for the treatable mimic, which most dementia diagnoses never are. The care from that point is real care, not a consolation: the memory services, the walking and bladder support, and the planning that families do better early than late. Some centers repeat or extend the test when the answer is borderline and the scan pattern is persuasive, so an unclear first result is a conversation, not a door closing. Either way, he leaves the process with a firmer diagnosis than he had.

Sources

Pymander is not a replacement for a physician and does not provide medical advice, diagnosis, or treatment.

Free AI doctor, 24/7 by textStart a free AI doctor consult