Normal Pressure Hydrocephalus (NPH) has remained a contested entity since its inception more than 5 decades ago. Recent articles have added fuel to the controversy surrounding this condition, including the coining of new terms such as “neurodegenerative NPH”. We sought the opinion of eminent experts, 3 Neurologists and a Neurosurgeon, who have contributed to the current discussion about NPH. We have asked them to answer specific questions and to formulate more generally their opinions on the subject.
Does normal pressure hydrocephalus exist?
Dr. Papageorgiou:
Although recently the entity of normal pressure hydrocephalus has been questioned, in everyday clinical practice it is evident that normal pressure hydrocephalus really exists. However, it is important to make a distinction between secondary cases (e.g. appearing after meningitis, subarachnoid hemorrhage or trauma) and “idiopathic” cases. Also, as the CSF pressure dynamics are abnormal (CSF pressure diagram during the day-night cycle), the term “chronic hydrocephalus of the elderly” which was proposed about 20 years ago could be a more accurate term.
Dr. Fasano:
Yes it does. It is not a disease, it is a complex and poorly understood syndrome, but saying that it does not exist is dangerous, as it would preclude many patients from an effective treatment. The debate about the existence of a condition reminds me of a similar situation in our field: we all know that Essential Tremor is a syndrome with many underlying causes and perhaps a common shared pathophysiology. Nevertheless and in spite of the many unknowns related to its biological underpinnings, we know how to approach patients with Essential Tremor from a practical standpoint.
Dr. Lang comment:
I agree with both opinions. The secondary cases are distinct and should be clearly differentiated from the “idiopathic” cases. “Idiopathic” NPH is a syndrome. To deny it exists would deprive patients from treatment that could have important impact on quality of life. This even applies to the patients that may have an underlying degenerative disease that might be recognized but still not contraindicate shunting with the hope of at least short-term benefit. However, “idiopathic” NPH is overdiagnosed and many patients received inappropriate neurosurgical therapy. From my perspective the critical issue is that this field needs a lot more research into every aspect from understanding pathogenesis and effective diagnosis to prediction of short and long term outcome.
What is the best criterion to decide whether a patient is a candidate for shunt placement?
Dr. Papageorgiou:
There is a number of clinical and imaging criteria useful for this decision. Although none of them can be considered as the best indicator, some combinations of criteria could be very helpful. For example, in the case of a patient with gait instability and urinary urgency, with a normal urinary examination and MRI findings typical for NPH, candidacy for shunt placement is straightforward. If one has to choose two of the best criteria, gait imbalance would be the best clinical criterion and “high convexity tightness” (high-tight sign) would be the best imaging criterion.
Dr. Fasano:
Patient has to have NPH, should have disabling symptoms and not present for too long. The real question is how to diagnose NPH with acceptable accuracy. We do not have perfect diagnostic tests and I rely on four main domains: 1. clinical features (and in particular a peculiar type of gait disorder started before cognitive/behavioral issues); 2. neuroimaging (and particularly the presence of DESH – disproportionately enlarged subarachnoid space – and/or the evidence of ventriculomegaly observed incidentally prior to the onset of neurological complaints); 3. a dramatic response to temporary CSF removal; 4. absence of signs clearly supporting the diagnosis of a degenerative conditions (e.g. supranuclear gaze palsy).
Dr. Lang comment:
Again, I agree with the comments provided. Important factors to consider are the possibility of long-standing arrested hydrocephalus that might have little to do with the clinical presentation, the most disabling symptoms and signs that are incompatible with the diagnosis and would not be expected to respond to shunting (e.g. very clear cortical features, such as pronounced language disturbances or apraxia), and appropriate symptoms that have been present for many years without much change (i.e. extremely chronic and stable) and that probably would not respond to surgical intervention (although we might still conduct the diagnostic testing, and if a surprising degree of benefit is seen, then the patient could be considered for surgery).
Is a lumbar drain or the removal of a large amount (how much?) of CSF via spinal tap the best approach for predicting response to shunting?
Dr. Papageorgiou:
Removal of a substantial amount of CSF (the “tap-test”) is still considered the “gold standard” for this decision. Imaging studies can also help in the diagnosis of NPH, but their predictive value for the effect of shunt placement is still not well-established. The amount of CSF to be removed has to be large (30-50 ml). However, in cases where it is technically difficult to remove a large amount of CSF, clinical improvement may still be seen with the removal of 20 ml of CSF.
A continuous lumbar drain is also an option because the production of CSF from the choroid plexuses is ~20 ml/hr, and thus, the amount of the CSF removed with the “tap-test” is rapidly replaced. However, the higher risk of infection makes the continuous lumbar drain a good choice only for cases in which the “spinal-tap” is negative (i.e. no improvement) but there is a strong clinical and imaging evidence of NPH.
Although the predictive response of the “tap-test” has been questioned, and a substantial number of false negative cases (that means that the tap-test was negative while patients improved after shunting) has been reported, the key question is whether the patient has been assessed sufficiently well before and after the spinal tap. A meticulous clinical assessment of the patient before and after the “tap-test” is of critical importance. This should include quantitative evaluation of:
Posture and Walking: number of steps and time to walk a pre-defined distance, balance tests (e.g. tandem walking of 3 meters recording time and errors), etc.
Cognitive tests with an emphasis on those that assess frontal executive functions and include time measures (i.e. TRAIL making test A and B)
Urinary urgency (time of urinary retention) and number of urinary losses in a certain time.
Finally, the time of the evaluation after the CSF removal is also important. This has to be done several hours after the CSF removal (~4 hs) and again after 24 and 48 hs. It is not infrequent to see patients having an improvement that disappears rapidly or patients having a delayed improvement.
Dr. Fasano:
Taking into account costs, safety and feasibility, I think that an outpatient-based lumbar puncture removing a large volume of CSF (35-40 cc), possible combined with gait analysis, is the best way to predict response. Unfortunately its sensitivity is not very high, reason why when I suspect NPH (based on the remaining three criteria already mentioned), I recommend admission for an external lumbar drainage. The sensitivity of the latter is better, but not 100%, and in rare case I recommend shunt surgery even when these tests are negative. Seldom I perform a sham tap test when I suspect an important placebo response.
Dr. Lang comment:
The obvious problem is that there is no diagnostic test that is 100% sensitive and specific in predicting the future response to shunting. Generally, I believe most investigators start with the “tap-test”, removing an adequate volume, and when this is negative but there is still a strong diagnostic suspicion typically a lumbar drain is the next step. I have heard some neurosurgeons argue that one can proceed without either of these approaches in “obvious” cases, however, I do not believe that this is a widely accepted opinion, except perhaps in very clear secondary cases. Although we reported the potential for a placebo response to large-volume tap, it has not been my practice to carry out sham tap tests routinely. On the other hand, I believe that the “tap-test” needs to be presented to the patient and family with equipoise, in a very neutral and un-leading fashion. Simply indicating that we would like to see whether there is any change, better or worse, to this intervention should be the approach rather than telling the patient that there is a tremendous potential to respond and that if there is benefit seen they would then go on to the “curative” neurosurgical procedure. The latter approach would be strong encouragement for a potential placebo response, depending on what the patient and family knows about the disorder and its treatment (e.g. from Dr. Google).
What is the best way to distinguish NPH from microvascular ischemic disease?
Dr. Papageorgiou:
This is a very difficult question. Co-morbidity of these conditions exists very frequently and is more frequent with advancing age. Vascular risk factors (mainly diabetes and hypertension), as well as hyperintensities on the MRI, are risk factors for NPH.Several clinical features (urinary urgency, frontal-release signs, cognitive impairment) are common in both conditions, although some differences exist. For example, pyramidal signs that are rare in NPH are frequently found in microvascular ischemic disease. Some imaging findings can be difficult to differentiate as well (e.g. periventricular hyperintensity due to microvascular disease or CSF leakage through the ependyma). However, the “high-tight” sign in the convexity, the marked dilatation of the Sylvian fissure, and in some cases the focal appearance of disproportionally enlarged subarachnoid spaces (DESH) are only seen in NPH.
Dr. Fasano:
There is no good way other than obtaining pathology. Since it is well established that NPH and microvascular ischemic changes are often coexistent, I do not focus on this differential diagnosis if I am sure about the diagnosis of NPH. The only exception is when patients are known to have significant vascular risk factors, as this can also affect the safety of our interventions. I should also say that often the hyperintensities seen with MRI are not ischemic but are rather the expression of transependymal CSF flow, which can be reversible in some cases after an effective shunting.
Dr. Lang comment:
Here, the important issue is to exclude patients with features that are incompatible with a diagnosis of “idiopathic” NPH. The clinical features, course of the disorder, accompanying imaging and response to CSF removal are particularly important here. We need much better clinical-imaging-neuropathological correlation of the imaging features that are often interpreted as microvascular ischemic disease. I agree with Alfonso that some of these changes may be related to transependymal CSF but there may be other pathological correlates (e.g. venous collagenosis) that require further study.
What is the best timing to proceed with shunt placement?
Dr. Papageorgiou:
One could say “the sooner, the better” because it seems that an early intervention can prevent cognitive impairment. However, in cases of advanced age (> 85) or patients with serious co-morbidities having a high-risk for surgical treatment (ventriculo-peritoneal shunting), repetitive removal of a large amount of CSF (e.g. every 3-6 months) may be preferable, at least initially. As with the most elderly patients, it is not infrequent to see a long-lasting improvement, even up to 6 months, and therefore, the repetitive CSF removal could be considered a good alternative treatment for selected cases.
Dr. Fasano:
As already said, there is a specific window as patients should have disabling symptoms but not present for too long, as shunting will not revert the white matter damage caused by a longstanding NPH. A common misconception is that clinicians should wait for the famous triad, which can lead to unneeded (and dangerous) delays. I focus on the disability caused by mobility problems. Please note that theoretically also NPH patients without disabling symptoms are candidates for shunt placement; however, due to the risks of the procedures, I usually monitor them closely so that we can act fast if symptoms become disabling. I have extensive discussions about the pros and cons of waiting with patients as well.
Dr. Lang comment:
Definitely the sooner the better in appropriately diagnosed patients with sufficient impact on quality of life that the surgical risk is justified. As much as we do not want to intervene too late, we should not be in a rush to operate on patients where the diagnosis is less certain or the risk outweighs the potential benefit.
Do neurodegenerative diseases associated with ventricular dilation (neurodegenerative NPH) improve with shunt placement?
Dr. Papageorgiou:
In the not rare case of a co-morbidity with a degenerative disease (~25%), a partial improvement is seen. In these cases, the improvement seen is related to the symptoms linked to NPH pathophysiology (e.g. improvement in gait and urinary continence without cognitive improvement, in a case with co-existing Alzheimer’s disease). This demonstrates the need to treat every case of suspected NPH, even in the presence of a neurodegenerative disease. However, it is obvious that in the case of a sole ventriculomegaly due to brain atrophy in a neurodegenerative disease, a clinical improvement cannot be expected. Thus, while a “tap-test” must be offered in every case of suspected NPH, even in the presence of a neurodegenerative disease, this suspicion should be supported by clinical and imaging evidence.
Dr. Fasano:
I am still puzzled with the concept of neurodegenerative NPH. It is still unclear whether these patients have the combination of two conditions (possible given the old age of these patients), if some neurodegenerative conditions can present with an NPH-like phenotype or if NPH predisposes to the occurrence of neurodegeneration (e.g. by affecting the clearance of pathological substrates). Simply speaking, I think that shunting can revert what is caused by NPH without affecting the neurodegenerative component. As long as these concepts are explained to patients and their families, I do not think that shunting should be precluded to these patients.
Dr. Lang comment:
I have definitely seen patients with the combination of a neurodegenerative disease and NPH. I have a suspicion that this occurs more frequently than expected, and therefore the neurodegeneration may very well somehow predispose to the CSF dynamic abnormalities of NPH. On the other hand, it is possible that the association is coincidental. The diagnostic approach should not differ between those thought to have an underlying neurodegenerative disease and those believed to have isolated “idiopathic” NPH. On the other hand, the discussion of whether to operate or not is definitely different because one can be more certain that the patient will have a worse long-term prognosis due to the underlying neurodegeneration. Thus, a detailed discussion and explanation to patients and families is critical. This is an area where further diagnostic testing with newer and upcoming biomarkers (e.g. biofluid (plasma, CSF), imaging) will almost certainly change the approach in the future.
Is the presence of a neurodegenerative condition an exclusion criterion for NPH?
Dr. Papageorgiou:
Of course not for the same reasons as explained above.
Dr. Fasano:
Yes if all the symptoms are attributable to the neurodegenerative condition, no if the patient has two different conditions.
Dr. Lang comment:
I agree entirely.
What is the single imaging feature most predictive of response to shunt placement?
Dr. Papageorgiou:
Maybe the “high-tight” convexity sign is the most significant positive sign of NPH having a predictive value of response to shunt placement. Periventricular hyperintensities around the frontal horns could be also predictive; however, their presence is not constant and they are often difficult to differentiate from those due to microvascular disease.
Dr. Fasano:
The presence of DESH followed by the evidence of ventriculomegaly observed incidentally prior to the onset of neurological complaints. In this respect, the relationship between ‘idiopathic’ NPH and ‘arrested hydrocephalus’ is not clear yet. The average head circumference of patients with ‘idiopathic’ NPH is greater than controls, thus contributing to the hypothesis that many of these patients lost the compensation that CNS put in place after an initial insult when they were very young. In neurology there are similar examples already (e.g. post-polio syndrome). For this and other reasons, I think that the differential between idiopathic and secondary NPH is artificial, perhaps with the only exception of the genetic forms of NPH (and interestingly a gene causing familiar NPH has been recently identified).
Dr. Lang comment:
Although Alfonso makes an interesting argument, I would still prefer to distinguish between the classical and overt secondary causes and the rather uncertain ones that might have caused arrested hydrocephalus with the insult occurring very early in life. Because ventriculomegaly is not uncommon as an ex vacuo phenomenon this is probably the least reliable of the classical imaging features. I agree that both the “high-tight” convexity sign and DESH are particularly useful.
Dr. Krauss comments:
Idiopathic normal pressure hydrocephalus (NPH) has remained a somehow enigmatic disease since its first description in the 1960s. While most clinicians agree that the clinical picture is characterized by a predominant gait disorder with the facultative presence of cognitive deficits and urinary urge or incontinence and the presence of “triventricular” hydrocephalus with tight parasagittal sulci, a detailed understanding of its development and its pathophysiology remains elusive. The term “idiopathic normal pressure hydrocephalus” has been recognized to be problematic since its introduction. First of all, the label “idiopathic” is a poor designation signifying that hydrocephalus would develop somehow “by itself” without having specific preconditions or being possibly causally linked to co-morbidities. This is clearly not the case given the proven relationship to vascular disorders and to arterial hypertension. Another problem is that we continue to use the adjective “normal” although it has become obvious over the decades that intracranial pressure is certainly not entirely normal in the majority of patients. It is great to see that, after many years of tranquilness and of limited interest, we now have discussions with controversies pointing at the very conceptual problems of the definition of this disorder.
The complexities of terminology and the lack of a full understanding of the etiology and the pathogenetic mechanisms of “idiopathic NPH” to my opinion should, however, not lead to the conclusion that the disorder per se does not really exist. The term “idiopathic NPH” might actually compromise a group of various disorders with different etiologies, and as such might either designate a clinical syndrome or a specific disease entity. Maybe it is time to think once more about reclassification of NPH and also about subcategorization. There might be several variants apart from the classical clinical and morphological picture. In particular, new classifications could consider more clearly the frequent comorbidities - may it be subcortical vascular encephalopathy, Alzheimer disease, Parkinson disease or PSP (which would be the “neurodegenerative” variant).
Two facts are often neglected regarding the surgical therapy of NPH. First, shunt surgery itself has become much safer - also along improved anaesthesiological regimes. Shunt surgery is straightforward and one of the most simple neurosurgical procedures. Shunt surgery can be offered also to very old patients and to patients with higher surgical risks, given the short operative time and the fact that patients can be mobilized actually on the day of surgery. To my experience, repetitive therapeutic lumbar punctures for CSF removal largely can be avoided. Second, shunt technology has improved remarkably over the last two decades. Programmable valves and gravitational “anti-siphon” devices allow handling of overdrainage or underdrainage problems with much more ease, and revision surgery is less frequently indicated than in the past. It is difficult to understand why some centers continue to use old technology with the argument to “reduce costs”.
Why not offer shunt surgery early at the onset of symptoms once the diagnosis has been confirmed? Of course, with that regard we have to discuss what actually compromises the “diagnosis of NPH” (see above). A clinical syndrome, a disease entity, a practical definition? Anyway, confirmation of improvement with a standardized protocol after high volume removal of CSF by lumbar puncture should be mandatory. Care should be taken to recognize strong placebo effects, and if results are negative, lumbar puncture can be repeated or lumbar drainage may be used for a couple of days. Note that the latter, however, might not be feasible in some settings. Finally, while it has been thought in the past that shunt surgery would provide benefit only for a period of several years, more recent studies with long-term follow-up have shown that there is stable improvement in many instances, which would also indicate that shunt surgery could be offered early, when symptoms are still reversible and when the progress of the disorder can be altered.