Introduction

In nearly a half century since idiopathic normal pressure hydrocephalus (iNPH) was first identified, the origins of its symptoms have remained obscure. A fundamental question is how ventricular enlargement and alterations in cerebrospinal fluid (CSF) dynamics in iNPH disturb the function of neurons in the brain. Most theories of pathogenesis have fallen short of fully explaining why iNPH is associated with the specific triad of gait, urinary, and cognitive disturbances rather than any other constellation of symptoms. There are also remaining uncertainties about the mechanisms responsible for the temporary remission of iNPH symptoms after external drainage of CSF and the more lasting improvement that follows shunt placement. Although many advances have been made in its diagnosis and treatment over the past 45 years, iNPH is still an enigmatic syndrome on many levels.

Given these limits to the present state of knowledge, it may seem quixotic to attempt to explain the pathophysiology of the cognitive deficits encountered in iNPH. However, the cognitive symptoms of iNPH have certain features that have made them amenable to study. The distinctive profile of cognitive deficits in iNPH provides clues to their localization in light of the brain’s established functional neuroanatomy. Structural and functional brain imaging have helped to elucidate the neural pathways involved in hydrocephalic cognitive impairment. In addition, studies of brain metabolism and blood flow, molecular imaging, and cerebrospinal fluid (CSF) physiology have provided novel windows into the pathogenesis of dementia in iNPH. While there is still much to be learned, there is considerable information available that is salient to understanding the origins of cognitive symptoms in iNPH.

Putative pathogenic mechanisms of iNPH

A number of pathophysiologic mechanisms have been identified that are potentially relevant to the pathogenesis of the cognitive symptoms of iNPH:

• 
  

  Mechanical distortion. Stretching of white matter and mechanical compression of subcortical gray matter caused by ventricular enlargement may be a cause of cerebral dysfunction in iNPH [1,2]. Progressive ventricular expansion distorts the corpus callosum and periventricular corona radiata [3]. It also compresses subcortical gray structures such as the hippocampus, caudate nucleus, and thalamus. The importance of subcortical mechanical distortion as a pathogenic mechanism has been questioned because cognitive impairment and other symptoms of iNPH can progress without gross expansion of the ventricles [1] and cognitive recovery can occur without visible reductions in ventricular size [4,5]. Mechanical distortion is therefore unlikely to be the sole cause of cognitive impairments in iNPH, but could well be a contributing factor.

  
  
• 
  

  Pressure effects. The modest increases in static intracranial pressure (ICP) of up to 240 mmH₂O that occur in iNPH are insufficient to cause neuronal dysfunction. Higher pressures can be recorded in iNPH patients as B waves (CSF pressure oscillations with a frequency of 0.5 to 2/min) that tend to occur transiently during sleep. B waves have not been directly linked to cognitive impairment nor would they be expected to cause the kind of persistent symptoms that iNPH patients typically experience. Increased CSF pulsatility (dynamic pressure changes that occur across the cardiac cycle) also occurs in iNPH and may secondarily impact on cognition through effects on cerebral perfusion and induction of hyperdynamic CSF flow.

  
  
• 
  

  Flow phenomenon. Hyperdynamic flow of CSF through the aqueduct of Sylvius is frequently observed in patients with iNPH [6]. The driving force for CSF flow is the systolic arterial pulsations that generate CSF in the choroid plexus and secondarily in the brain parenchyma. During systole, there is outward radial movement of the ventricular walls followed by flow of CSF through the cerebral aqueduct. However, owing to ventricular enlargement and decreased elasticity of the brain in iNPH, there is a reduced capacity for radial ventricular wall movement. This energy of the arterial pulsation is instead converted into a translational force driving CSF through the aqueduct at a greater than normal velocity and a hydrodynamic force driving transependymal CSF resorption. The rapid downward movement of CSF flow can exert a hammer-like effect on the roof of the brainstem causing damage to the nuclei such as the substantia nigra (SN). Nigral depigmentation and neuronal degeneration can result in parkinsonism as well as cognitive impairment. Since not all iNPH patients exhibit parkinsonism, this mechanism is most likely not a universal cause for iNPH symptoms but may be a factor in some cases.

  
  
• 
  

  Toxin accumulation. CSF production has been reported to be reduced in iNPH when measured by bulk flow protocols such as the Masserman technique [7]. Since the circulation of CSF plays an important role in removal of the byproducts of brain metabolism, decreased CSF production in iNPH could lead to accumulation of toxic metabolites that cause neuronal dysfunction. Decreased clearance of beta amyloid in this context may be a factor in the association between iNPH and Alzheimer’s disease (AD). It has been suggested that cerebral metabolic derangement in iNPH begins in conjunction with derangements in CSF dynamics but eventually becomes uncoupled and self-sustaining.

  
  
• 
  

  Neurochemical disturbances. Disturbances in dopaminergic, glutaminergic, GABAergic, cholinergic, and serotoninergic neurotransmission have been found in kaolin models of hydrocephalus and may also be associated with iNPH (see Kondziella et al. [8]). Unlike AD, in which selective degeneration in the cholinergic system is a relatively early manifestation of neurodegeneration, neurochemical deficiencies are more likely late effects of iNPH and reflect compromise to the integrity of structures juxtaposed to the CSF spaces, such as the substantia nigra, striatum, thalamus, and hippocampus as well as their white matter projections.

  
  
• 
  

  Interstitial edema. Interstitial edema occurs in the brain parenchyma in iNPH. This fluid accumulation is often visible in the periventricular regions on brain MRIs as T1 hyperintensities or T2 hypointensities that can often be reversed by shunt placement. The occurrence of interstitial edema may be a consequence of increased transependymal flow of CSF from the ventricles into the brain parenchyma and/or hydrostatic forces within the brain related to increased cerebrovascular resistance. Cerebral edema is a known cause of cerebral dysfunction and it represents a likely factor in the development of cognitive deficits and other neurological symptoms in iNPH. In experimental models, interstitial edema has been shown to decrease cerebrovascular reserve and cause other disturbances in regulation of cerebral blood flow [9]. These vascular mechanisms are also potential precipitant factors for iNPH’s cognitive symptoms (see below). Successful shunt placement can result in reduction in interstitial edema that parallels improvements in iNPH symptoms.

  
  
• 
  

  Cerebrovascular compromise. Cerebral blood flow (CBF) studies in iNPH have found global decreases but no clear correlation between resting perfusion and response to shunt [10,11]. Focal decreases in CBF involving mesial frontal and anterior temporal regions have also been documented with some correlation to shunt responsiveness [12]. While decreases in resting CBF do not seem sufficient to explain symptoms of iNPH, decreased cerebrovascular reserve (CVR), as can be documented by a diminished response to an acetazolamide challenge test, could contribute to inadequate cerebral perfusion during periods of heightened brain activity. Possible causes of altered CBF and CVR in iNPH include all of the above-mentioned mechanisms (mechanical distortion, pulsatile pressure increases, toxins, or interstitial edema) as well as other factors that compromise cerebral blood flow.

It seems likely that a combination of these and possibly other mechanisms are contributing factors in the pathogenesis of cognitive symptoms in iNPH. A possible synthesis of these mechanisms would be that an imbalance of CSF production and clearance leads to progressive ventricular enlargement. Resultant increases in CSF pulsatility and reflux of CSF into the ventricle drive the roof of the lateral ventricles upward, trapping the sagittal sinus between the hemispheric convexities and the skull. This further compromises CSF clearance and fosters the generation of alternative pathways for CSF movement through the brain parenchyma, particularly by transependymal resorption. As CSF moves across the ventricular ependyma into the interstitial spaces of the brain, interstitial edema develops in periventricular structures. This edema compounds the compressive effects of ventricular expansion and leads to reductions in cerebral perfusion and cerebrovascular reserve. Diminished CSF clearance may also foster accumulation of toxic metabolites which further disturb cerebrovascular autoregulation and cause neuronal dysfunction and secondary neurochemical disturbances. Unless iNPH is treated, vascular insufficiency eventually causes periventricular white matter infarction, at which point the symptoms become irreversible.

Based on the most commonly observed pattern of interstitial edema in cases of iNPH, it can be hypothesized that the early cognitive symptoms of iNPH reflect the development of interstitial edema and vascular compromise in the white matter pathways that pass near the anterior and posterior horns of the lateral ventricle. These regions are mechanically prone to reverse transependymal fluid movements which appear as ventricular “capping” on MR images. Gray matter structures juxtaposed to the ventricles, such as the caudate, thalamus, and hippocampus, may also be affected. The cognitive symptoms would therefore be considered manifestations of pathophysiological processes that are relatively circumscribed spatially and predominantly subcortical in nature.

Involvement of the cerebral cortex in the symptoms of iNPH is suggested by the results of some metabolic and perfusion imaging studies. Decreased cerebral blood flow and/or metabolism documented by imaging in the cortex of iNPH patients can be reversed by shunt treatment [13,14]. However, these imaging studies do not distinguish between a primary disturbance within the cerebral cortex and the distal effects of dysfunction (diaschisis) in the white matter tracts connecting to these regions. Diaschisis seems a more likely explanation since postmortem studies have failed to provide direct evidence of cortical pathology in iNPH.

General features of cognitive impairment in iNPH

The profile of cognitive impairments in iNPH is recognizably that of a subcortical pathological process. In this respect, iNPH resembles Huntington’s or Parkinson’s disease in which dysfunction in subcortical frontostriatal circuits leads to impairments of complex attention, speed of information processing, phonemic verbal fluency, and executive functions. In contrast, the pathology of Alzheimer’s disease more directly involves neocortical areas, resulting in disturbances of semantic language, calculation abilities, and other cortical functions that are relatively spared or only secondarily affected in iNPH.

The extent of impairment in cognition in iNPH can range from subtle neuropsychological dysfunction to a frank dementia [15]. Early cognitive symptoms can readily go undetected in iNPH or can mistakenly be attributed to normal aging. In our experience, many high functioning patients do not report subjective cognitive changes early in the disease course. While these patients perform well on global screening measures (e.g. MMSE), they frequently exhibit some form of cognitive compromise when more detailed neuropsychological testing is performed. Therefore, even when the patient denies cognitive symptoms, formal cognitive evaluation is recommended. This should be pursued in any patient with suspected iNPH, even in patients who perform well on global screening measures during neurological work-up. The components of a neuropsychological battery useful for serially assessing patients with suspected or confirmed iNPH are presented in Table 7.1 [16–25].

Table 7.1 Sample neuropsychological battery for serial assessments in iNPH

Note: * measures repeated to assess change

In the early stages the presentation can include mild decrements in psychomotor speed, reduced information processing, and inefficiency in aspects of learning and memory [15,26–28]. Reaction time and psychomotor speed are also typically impaired at early stages. The cognitive symptoms of iNPH are more likely to be reversible when the disease is in its early stages.

Over the course of several months to years cognitive symptoms of iNPH typically progress. As a result, patients may show more severe reductions in psychomotor speed, information processing, executive functioning, complex attention, and memory. In later stages, speech output may be reduced secondary either to dysexecutive syndrome or motivational problems. Pragmatic aspects of language may also be affected.

With progression of disease in the absence of treatment, cognitive and other symptoms of iNPH can become intractable. Delays in diagnosis and/or treatment can lead to the development of a syndrome that strongly resembles the late stages of Binswanger’s disease [29]. The occurrence of confluent periventricular ischemic changes in untreated and late stage iNPH suggests that cerebrovascular compromise is one of the key mechanisms underlying the cognitive symptoms of iNPH.

Changes in gait impairment are commonly used to monitor improvements after drainage of CSF and shunt placement in cases of iNPH. While gait symptoms can provide sensitive measures of response, cognitive symptoms are also amenable to treatment. Specific tests that measure arousal, affect, and cognition before and after drainage of CSF have been shown to be a useful adjunct to gait testing in assessing candidacy for shunt [30]. Improvements in cognition can occur very rapidly after shunt placement and these gains may increase over the subsequent weeks to months. Evidence from serial phase-contrast MRI studies [31] suggests that iNPH symptoms may improve as aqueductal CSF flow normalizes after shunt placement, a process that can take several weeks with a fixed pressure shunt valve and longer with a programmable valve that requires serial adjustments.

Neuropsychological evaluation in iNPH

The neuropsychological examination is an extended mental status evaluation that assesses a broad range of cognitive skills and employs measures that are more sensitive to early changes than global cognitive screening measures. The general domains that are assessed as part of a neuropsychological evaluation include attention, memory, executive functions, visuospatial skills, psychomotor and motor skills, mood and other aspects of psychological functioning. The length of an evaluation varies based on the specific questions that are being asked. A typical battery that is focused on differential diagnosis of dementia ranges from 1.5 to 3 hours. The battery may be tailored slightly for iNPH patients to include a greater number of motor and psychomotor tasks, since these skills will likely be reassessed as part of the diagnostic process and as a gauge of outcome if patients undergo treatment. The initial evaluation serves both as the differential diagnostic work-up and as a means to establish a baseline level of functioning. In specialty centers, post-drainage testing is also performed. These evaluations are typically much shorter and focus on a subset of baseline tests that are helpful in prognosticating regarding candidacy for treatment. Patients who undergo treatment will also be seen for follow-up evaluation at regular intervals to track improvement and monitor response to treatment. In a clinical setting, the frequency and length of the follow-up evaluation will depend on the particular circumstances and questions that need to be addressed. A sample baseline neuropsychological battery is presented in Table 7.1 and measures that are repeated with diagnostic testing (e.g. drainage) and after treatment are noted.

Localization of neuropsychological deficits in iNPH

Response to CSF drainage (tap test, external lumbar drainage, shunt)

While iNPH has been described as a reversible form of dementia, for many years, little was known about the recovery of cognitive functions following treatment. Quite a few studies had documented improvement in overall mental status (see Klinge [43]), but there was a lack of agreement about whether specific cognitive deficits responded to treatment. Other studies have found no improvement in cognition following shunting [26,44]. There was a notion set forth that more globally demented patients showed the most dramatic recovery following shunt placement [15], with less affected patients being less likely to improve cognitively after shunt. However, recent studies employing more detailed cognitive test batteries have not supported this finding [30,32,33,38].

In recent years there has been an increasing interest in studying outcome in iNPH, and several well-designed studies have clearly documented that many aspects of the cognitive syndrome can be improved with successful treatment [30,32,33,38]. There have also been investigations looking at initial response following CSF removal as a way to predict who will respond to shunt placement. The mechanisms that underlie the immediate, early improvement in motor and in some psychomotor tasks are not well understood. Tsakanikas et al. [45] demonstrated that tests of upper extremity motor functions improved following tap test and may be utilized to predict response to shunt. A recent study examining response to external lumbar drainage [5] showed clinical improvement in gait and also in psychomotor speed and attention in 50% of the patients. Interestingly, these clinical gains were not associated with changes in ventricular volume. These authors point to an “inverted hysteresis” as an explanation for the occasional phenomenon of prolonged improvement following drainage, suggesting that the alterations in neuronal integrity seen following CSF removal endure even after the CSF is replaced, “before the pathological mechanism gets control again.”

There is mounting evidence that there are real and lasting improvements in cognition in carefully selected iNPH cases that undergo shunt placement. Duinkerke et al. [38] examined cognition following shunt placement in 10 iNPH patients and found that 6 of the 10 patients showed improvement in at least 50% of the tests administered, with the greatest improvement seen in verbal memory for both word lists and contextual information (story memory). Although a number of early investigations found no improvement in complex executive functions following treatment [15,26], multiple recent studies have shown clear gains in aspects of executive skills following treatment. Hellström et al. [32] showed an improvement in performance on the Stroop test following shunt, suggesting gains in attention and inhibitory control. Katzen et al. [30] showed that after controlling for practice effects in a small iNPH cohort, improvements were observed following shunt in Trail Making B, a test of complex attention and set shifting. A trend was also observed for symbol digit modalities (SDMT). The post-shunt performance at 6 months was improved for several of the other measures of processing speed and motor functions, but the differences were not significant after controlling for the practice effects in the control group in this small sample. Others have argued that there may not be an expected practice effect in iNPH or that practice effects may only be observed on certain measures [46]. The use of reliable change indices to address these issues has been suggested.

Several other investigators have also shown improvement in Trail Making B, Trail Making A, and SDMT after treatment [13,38,39]. One study of 185 iNPH patients conducted in Spain found that performance on Trail Making B was unchanged despite observed improvements in Trail Making A [33]. In that investigation, less than half of the subjects completed Trail Making B at baseline (n = 70) and only 13 additional patients performed the test at follow-up (n = 83), which suggests that this test may be too difficult for severely impaired patients. The presence of a floor effect on Trail Making B makes it challenging to assess changes following treatment in more impaired samples. Modifications to test administration and scoring procedure may be required to adequately capture performance gains in NPH. Further, additional measures of psychomotor and executive skills should be developed in order to better capture these changes.

Few investigations have examined changes in visuospatial skills and visual memory. One study that demonstrated baseline visuospatial impairment did not show improvements in these skills after successful treatment [39]. With regard to visual memory, the results are mixed with some studies showing that visual memory improves with shunting [33] and others not showing any observable gains in this cognitive domain [38].

Motor skills including simple motor speed, fine motor coordination, and motor precision [13,32,33] have been shown to improve following treatment. However, when comparing change in performance to a group of controls, the gains were no longer statistically significant in one study [30]. As with the findings for some of the executive and attentional measures, this may be secondary to practice effects, but larger cohorts are needed to examine whether there are in fact true practice effects in an NPH sample. Improved methods for determining what constitutes an improvement are also warranted (e.g. reliable change indices).

It is notable that while there is growing evidence that well-selected iNPH patients show improvement across many cognitive domains following successful treatment, cognition does not appear to return to “normal” in most cases. Hellström et al. [47]showed widespread cognitive improvement; however, the performance of the iNPH group remained below the performance of a healthy comparison group, suggesting that the recovery was incomplete.

Differential diagnosis/comorbidities

Variability in the cognitive presentation of iNPH is in part due to the heterogeneity of the disease, but also reflects variability in the latency for a diagnosis to be made. While some patients are identified swiftly via clinical work-up and brain imaging, others do not present to medical attention for months or even years after symptoms begin. In fact, two recent studies examining outcome in iNPH patients both showed that the average duration of symptoms was 24 months (range 12–36 months) [32,33]. Both studies examined whether duration of symptoms was a factor in outcome. One investigation demonstrated that shorter duration was in fact associated with global cognitive improvement and improvement in motor speed [33], whereas the other study did not find a similar association [32]. Despite the uncertainty regarding how duration of symptoms impacts outcome, delays in presentation clearly make differential diagnosis more challenging and the presence of comorbidities further complicates the presentation.

It is not uncommon, especially if diagnosis is delayed, for the neuropsychological test results to suggest involvement of not only frontal systems, but a more widespread cognitive decline that may indicate comorbid AD, vascular cognitive impairment, or another neurodegenerative process. While the existence of cortical deficits such as aphasia, agnosia, and alexia early in the presentation are highly suggestive of an alternate diagnosis, it should be noted that these deficits can be seen in the more advanced stages of iNPH. In some cases they indicate a greater likelihood that a comorbid neurodegenerative disease is also present.

Differential diagnostic considerations

It is important to point out that the presence of a comorbid disorder does not negate the fact that iNPH may contribute to the presentation, and even more importantly, is not necessarily a contraindication for treatment. In our experience, many patients with comorbid neurodegenerative conditions have been successfully treated for iNPH. While the cognitive symptoms typically do not show substantial improvements post-shunt in patients with significant comorbid disease, improvements in gait can be associated with increased independence in activities of daily living and can significantly improve the patient’s quality of life as well as making physical management easier for the caregiver.

As described above, there is great variability in the cognitive presentation of patients with iNPH. The profile of deficits typically follows a frontal subcortical pattern, which can often be difficult to distinguish from other disorders that interrupt the integrity of frontal subcortical pathways such as vascular etiologies, parkinsonian syndromes, and even pseudodementia.

Differentiating iNPH from other dementia syndromes is particularly important given that the signs and symptoms can be at least partially reversed with the diversion of CSF through shunt placement. It has been suggested that iNPH may be a contributing factor in to up to 6% of dementia cases [48]. Given the challenge of identifying and differentiating iNPH, this number may be an underestimate. Further, this figure focuses on iNPH cases in which the cognitive presentation has advanced to dementia, yet there are numerous circumstances where mild cognitive dysfunction is a presenting feature and even subtle cognitive changes can impact quality of life and one’s ability to function optimally both at home and at work.

Because Alzheimer’s disease is so prevalent in the age group at risk for iNPH, distinguishing cognitive symptoms of iNPH from AD is the most frequent differential diagnostic question. There are some data comparing profiles of iNPH and AD groups. While the overlap in cognitive presentation with these disorders may give us clues into the underlying neuropathological mechanisms, additional work is needed in this area to better understand the pathophysiology of iNPH and the mechanisms that are involved both in progression to a dementia that is indistinguishable from AD and the reversibility of certain cognitive features.

Conclusions

Cognitive symptoms of iNPH reflect disturbances in spatially disparate brain networks that mediate executive function, attention, memory, motor sequencing, and visuospatial perception. Cognitive dysfunction in iNPH reflects diverse pathophysiologic processes that are most pronounced in subcortical regions, especially the periventricular white matter of the corona radiata and corpus callosum and subcortical gray structures such as the striatum, thalamus, substantia nigra, and hippocampus. Cognitive disturbances in early stages of iNPH are typically weighted towards frontal systems and are thought to be the product of reversible processes such as interstitial edema and diminished cerebrovascular reserve. Without treatment, iNPH’s cognitive symptoms progress and can become irreversible as a result of infarction of white matter and associated metabolic derangements within the brain. With careful selection of shunt candidates as well as management that avoids surgical and postoperative complications, cognitive dysfunction in NPH can be reversed, although only rarely to pre-morbid levels. Common neurodegenerative disorders in the elderly such as AD can coexist with iNPH and diminish the prognosis for long-term recovery of cognitive function after shunt. However, shunt placement can ameliorate symptoms even in cases with coexisting iNPH and neurodegenerative pathology.

Noninvasive methods of studying the structure and function of the human brain and cerebrospinal fluid dynamic are rapidly improving. Ongoing research can be expected to shed further light on the relationship between altered CSF dynamics and neuronal dysfunction leading to cognitive disturbances in iNPH. This should provide new avenues to diagnosis and treatment of iNPH, including the future prospect of improved surgical management and pharmacologic treatment and, ultimately, prevention of the disease.