In this issue of the AJNR, Bradley et al postulate that normal pressure hydrocephalus (NPH) begins in infancy as benign external hydrocephalus due to impaired drainage of CSF by the arachnoid villi and granulations. As the cranial sutures allow for calvarial vault expansion in infants with this condition, their heads are larger than normal. To support their hypothesis, the authors measured intracranial volume in patients with suspected NPH and found the volumes to be larger than with age-matched control subjects. Their speculation is unlikely to ever be proved, for it would require imaging studies from a large number of normal infants to determine the incidence of benign external hydrocephalus, a confirmed relationship between benign external hydrocephalus to intracranial volume, and a follow-up period spanning more than 50 years to correlate the findings in infancy with those individuals who subsequently develop NPH. Benign external hydrocephalus may be under diagnosed, because most infants do not undergo CT or MR imaging studies. It is more common to see benign external hydrocephalus associated with big heads, because an infant whose head circumference is at or above the 98th percentile will more likely undergo imaging. In most cases, the infant has a big head on an inherited basis, because large heads are often present in the family. There also appears to be a correlation between body size and head size. The authors of this study did not have any data relating body size to head size.

The diagnosis of NPH still remains difficult. Most would agree that a patient older than 60 years with progressive dementia, ataxia, and incontinence with large ventricles who responds to CSF diversion most likely has NPH. Often the ventricular size before the development of symptoms consistent with NPH is not known. With idiopathic NPH, no known cause is evident, whereas secondary NPH results from subarachnoid hemorrhage, meningitis, trauma, and so forth. In retrospect, all NPH may be secondary; the idiopathic form is of unknown etiology. It would appear, however, that both forms of NPH are a result of increased resistance to CSF drainage.

The authors selected 51 patients that met their criteria for inclusion in the study. Only MR imaging data were used to make a diagnosis. Because no clinical data were supplied, we do not know how many fulfilled the criteria of having NPH preoperatively nor do we know whether these patients responded to CSF diversion. Although the intracranial volume was significantly larger in the patient population studied, there is no guarantee that the next 51 patients studied for NPH would necessarily have a significantly increased intracranial volume. Thus, this finding may just be a statistical fluke akin to flipping a coin and getting 10 heads in a row.

The authors postulate that benign external hydrocephalus is secondary to decreased absorption from immature arachnoid villi. The observation that several laboratory animals and infants have arachnoid villi, but no arachnoid granulations, was cited by Dandy and Blackfan (1) as an argument against these structures having an important role in CSF absorption. Subsequent investigations have shown that the arachnoid villi and granulations are anatomically and functionally the same, the only difference being that the arachnoid villi are not visible to the unaided eye. If the arachnoid villi are the major site of CSF absorption, their numbers and individual structures should have some bearing on the development of hydrocephalus, including the benign external variety. That the number and size of arachnoid granulations increase with age need not have any implications, because there does not appear to be any correlation with size and number of granulations and their ability to absorb CSF. If a significant fraction of CSF and its constituents drain via the lymphatics, what is occurring at the arachnoid villus is less relevant (2–4). Regardless of whether the CSF is being drained by the villi or the lymphatics, a generous subarachnoid space could indicate increased resistance to CSF absorption. As the name implies, however, the situation is benign, as the amount of CSF over the convexities of the brain does not progressively increase (in fact, the CSF volume on the brain surface diminishes as the infants age) nor do the ventricles appear progressively larger on follow-up studies in most of these infants. This finding also raises the question as to why in benign external hydrocephalus, the volume of fluid in the subarachnoid spaces increases but not in the ventricle, whereas in the nonbenign form of progressive hydrocephalus, it is the ventricles that enlarge.

The extracellular space in the parenchyma is approximately 15%. There is a continuous movement of fluid in the extracellular space either toward the ventricles or subarachnoid space, depending upon concentration and osmotic gradients, as there is no barrier at the ependymal or pial surfaces. Water in and of itself is freely diffusible across the blood vessels in the parenchyma. Because the only known force responsible for bulk CSF absorption is a pressure gradient, if the pressure on the outside of the capillary is higher than inside, it would cause the capillary to collapse and prevent any absorption. Thus, there is no bulk absorption of CSF from the parenchyma. This does not, however, mitigate free water exchange. A good example of this is the use of an osmotic diuretic such as mannitol to shrink the brain. The attenuation or signal intensity changes in the periventricular region seen on CT and MR imaging studies, respectively, in acute, but not chronic hydrocephalus, is indicative of migration of fluid but does not equate with absorption.

As the authors note, there are often changes seen in the deep white matter in patients with NPH on MR imaging studies. The idea that age-related increased resistance to CSF flow through the parenchyma or changes in elastance might be a factor in the development of idiopathic NPH is an interesting concept to explore.

NPH is indeed a treatable disease with CSF diversion. The problem is trying to diagnose the disease in those patients with large ventricles who would benefit from shunt surgery. The main concern is the high complication rate associated with CSF diversion. If one could markedly reduce shunt complications, then the criteria for shunt surgery could be liberalized as a decrease in shunt complications would alter the risk-benefit ratio.

• Copyright © American Society of Neuroradiology