By Brian Pikul, MD

Almost four decades have passed since the term Normal Pressure Hydrocephalus (NPH) was first used to describe a triad of symptoms, concurrent cerebral ventriculomegaly (communicating hydrocephalus) and normal pressure on a random lumbar puncture. These patients develop a set of symptoms that are hallmarked by one or more of the following: gait instability, urinary incontinence and dementia. The ventriculomegaly is cerebral ventricular enlargement where measurements of intracranial pressure are found to be within the normal range. The idiopathic form of NPH is the most common. The brain has adapted to the cerebrospinal fluid stagnation and has not experienced a buildup of high intracranial pressure. Its onset is typically between 60 and 90 years of age. A secondary variant is believed to be caused by conditions such as subarachnoid hemorrhage (SAH), central nervous system (CNS) infections/meningitis, trauma, posterior fossa operations, brain tumors, aqueductal stenosis, and other conditions. These secondary conditions can occur, and therefore cause NPH at any age.

Gait disturbance is the most frequent NPH symptom. It occurs in 90% of these patients. An initial balance disturbance is often followed by a "short-stepped" and wide-based gait, and difficulty with making turns. The gait is often referred to as "magnetic" as the patient appears to take short steps where the foot is somewhat forcefully planted on the floor. There do not appear to be lateralizing cerebellar signs.

Cognitive changes occur in 80% of patients with NPH. Reaction time is increased and there is a slowness of thought. Dementia with short-term memory loss is a prominent feature. However, as we will see, the timing of the onset of these cognitive changes carries with it a predictive ability regarding response to treatment. There is often difficulty in diagnosing the difference between other neurodegenerative conditions causing dementia and those associated with NPH.

Urinary incontinence occurs in 45-90% of NPH patients. The release of urine is usually unknown to the patient.

There is no papilledema. As stated above, spinal tap leads to discovery of normal CNS pressure.

No diagnostic test has led to good predictive response to CSF shunting, the current mainstay of treatment. Likewise, the classic triad of symptoms leads to a 65% positive predictive value to shunting.

Lumbar puncture should reveal an opening pressure of less than 20cm H20. Large volume lumbar puncture (40-50cc) may allow release of enough CSF to lead to temporary improvement lasting 1 day to several weeks. Lumbar CSF drainage for the period of several days (3-5 days) may demonstrate improvement in patients who fail to show improvement after a single large volume lumbar puncture tap. The longer period of temporary external lumbar CSF drainage leads to greater sensitivity and specificity in predicting response to shunting.

Radionuclide cisternography as a predictor of improvement after shunting is controversial. Persistent ventricular radionuclide activity at 48-72 hours post-injection may be the only nuclear medicine predictor of shunt outcome.

Head CT scanning demonstrates ventriculomegaly. The degree of ventricular dilatation by measurement of ventricular size does not correlate with response to shunting. Patients with cortical atrophy may respond to shunting between 50% and 60% of the time. Therefore, compression of the sulci over the convexity is not necessarily a predictor of outcome. Post-shunting changes in ventricular size are not required to see symptom improvement.

Head CT and Brain MRI scanning can detect periventricular lucencies (low densities) that may represent trans-ependymal CSF flow abnormality.

Brain MRI may help diagnose other causes and is promising for predicting outcome. Small studies have shown correlation with symptom improvement by evaluation of CSF dynamics.

Single Photon Emission Computed Tomography (SPECT) to evaluate blood flow has only predicted positive response to shunting when subcortical low flow areas are enlarged. PET (Positron emitted tomography) has not correlated with shunt outcome.

Diversion of CSF from brain to either peritoneum, venous vascular system or pleura via a small silastic conduit with a valve pressure regulator remains the standard of care for patients with the triad and supporting imaging/CSF pressure findings. The response of patient symptoms to shunting is approximately 60% after shunt placement and durable in its response in one-half of those patients. Shunt success in alleviating symptoms is the best current method to diagnose the condition. Gait and urinary continence improvement are the hallmarks of response to shunting. Patients who do not have gait disturbance as part of their constellation of symptoms have a poor response to shunting. Likewise, patients who have dementia as the initial symptom also do not have a positive response to shunting. Although dementia can improve with shunting, it is the one symptom of the triad that does not improve significantly if it is severe at the time of surgery. Unfortunately, complications such as over- or under-drainage of CSF can occur up to 30% of the time in published series. Newer programmable shunt valves hold promise for reducing the need for shunt revision surgery as pressure changes to the valve can be made externally. This is performed with a magnetic or a computer programming device in much the same way that a pacemaker is reprogrammed externally.

In conclusion, the response of NPH to CSF diversion surgery (i.e. shunting) is best in patients who have the complete triad of symptoms, gait instability, urinary incontinence and dementia, and have had early gait disturbance. Prolonged pre-operative CSF drainage and imaging findings of low blood flow in the subcortical regions correlate with symptom improvement after shunting. The amount of hydrocephalus does not correlate with symptom improvement. Some 60% of patients improve after shunting with one-half of these improvements found to be durable over the long-term.