Background

Normal pressure hydrocephalus (NPH) is a clinical syndrome characterised by unsteadiness of gait, urinary incontinence and cognitive decline, which can respond to cerebrospinal fluid (CSF) drainage [1]. The radiological “hallmark” of NPH is markedly enlarged ventricles in the absence of involutional change in the brain. NPH has been categorized as either idiopathic (iNPH) or secondary (following head trauma, intracranial haemorrhage, tumour, meningitis or complication of cranial surgery).

iNPH remains a controversial entity, and of the relatively few patients coming to autopsy, a significant proportion have been noted to have co-existing neurodegenerative or vascular pathologies [2, 3]. Whether those pathologies are the primary cause of the syndrome or an additional incidental disease process in the elderly population affected is uncertain.

Two recent reviews have reported the post-mortem findings of 38 patients that had been clinically diagnosed with iNPH with vascular changes (71 %), Alzheimer’s disease (AD) (61 %), progressive supranuclear palsy (PSP) (2.6 %) and corticobasal degeneration (CBD) (2.6 %) being the most frequent co-existing pathologies. None of these patients had iNPH as the only final clinical diagnosis [2, 3] and no specific neuropathological changes for iNPH were identified [5]. Some of the patients had an unsuccessful shunt procedure, raising questions about the accuracy of diagnosis in these particular cases.

In a prospective study of 37 patients who underwent shunting for iNPH, a high prevalence of AD pathology was found in cortical biopsies obtained during shunt insertion. Patients with moderate to severe pathology failed to demonstrate benefit on any study measures assessing gait, cognition and bladder control at 4 months follow-up [4].

Hydrocephalus in elderly rats can induce both amyloid beta (Aβ) and phosphorylated tau accumulation in a pattern similar to human AD. CSF dynamics also appeared to impair Aβ clearance as a result of impaired CSF circulation [5].

In the present study, we aimed to review the clinical, radiological and pathological characteristics of four patients who had received a clinical diagnosis of iNPH.

Methods

We reviewed the database of 1,450 brains in the Queen Square Brain Bank (QSBB). The retention of and access to clinical records and pathological material at QSBB has generic ethical approval from the London Multi-Centre Research Ethics Committee.

Our inclusion criteria were patients aged 50 years or above exhibiting: (1) presence of at least two of the following symptoms or signs (a) a progressive gait impairment, (b) severe urinary difficulties, (c) cognitive decline; and (2) markedly enlarged cerebral ventricles on computer tomography (CT) or magnetic resonance imaging (MRI) of the head who underwent an invasive diagnostic (high volume lumbar puncture) or therapeutic procedure (insertion of a ventriculo-peritoneal [VP] shunt) [6] for the evaluation or treatment of suspected NPH. Patients with a history of congenital hydrocephalus, venous sinus thrombosis, central nervous system infections or obstructive lesions were excluded.

A systematic review of all available medical records was performed. Case notes were reviewed by one of the authors (NKM). The neuropathological diagnosis of all cases was made by a neuropathologist at the QSBB (JH or TR). Case 3 has been previously published [7].

Results

In four cases (0.3 %), a clinical diagnosis of NPH was made by either a neurosurgeon or neurologist. All patients had been assessed during life by at least one neurologist, and three out of four by a movement disorders specialist. Their demographic and clinical features are summarized in Table 1.

Table 1 Table showing clinical characteristics of four patients initially diagnosed with normal pressure hydrocephalus
Full size table

All four patients presented with difficulty walking and falls, postural instability, and mild subcortical or executive dysfunction. Three patients had mild difficulties with swallowing. Case 4 had urinary urgency. Two had risk factors for secondary NPH (Table 1).

Neurological findings of all patients prior to the diagnosis of NPH revealed a slow, broad based and unsteady gait with minimal limb rigidity. None of the patients had tremor or an eye movement abnormality. The Mini-Mental Score Examination (MMSE) ranged from 26 to 28 out of 30.

On neuroimaging (two patients had CTs and two had MRIs), all patients had ventricular dilatation disproportionate to the degree of cortical atrophy on visual assessment. In the second MRI of case 4 (6 years into the disease) midbrain atrophy and the ‘hummingbird’ sign was also noted.

Three patients had transient improvement in 10-m walk and mental tasks (cases 1, 2 and 3) after 13–40 mls of CSF withdrawal. CSF opening pressure was 13 cmH2O in case 1 and 21 cmH2O in case 3. It was not recorded in case 2 as lumbar puncture was done with patient seated. case 4 did not undergo trial of CSF withdrawal prior to VP shunting as gait dysfunction, cognitive decline, urinary disturbance and imaging studies were thought to be highly typical of NPH.

Two patients underwent VP shunting with transient improvement in gait and cognition. In the other two patients (cases 1 and 2), VP shunting was suggested by neurosurgeons but the decision was made by neurologists not to proceed, as they felt there was an underlying co-existing neurodegenerative condition.

All four patients had striking gait instability, falls and freezing initially. Progressive cognitive decline, significant urinary disturbance, eye movement abnormalities, hypomimia and hypophonia occurred later in the disease course. These later features prompted the neurologists that shunt-responsive NPH was not the underlying diagnosis.

At post-mortem, case 1 was found to have Parkinson’s disease (PD) and the other three patients had pathologies consistent with PSP. All patients had additional mild to moderate pathological ageing changes and three patients had evidence of small vessel disease. Figure 1 shows moderate dilatation of left lateral ventricle with frontal atrophy in case 4.

Fig. 1
figure 1

Case 4 (pathological diagnosis: PSP) showing moderate dilatation of left lateral ventricle with frontal atrophy

Full size image

Discussion

We report four patients in whom an initial diagnosis of iNPH was considered and who had at least a transient response to large volume CSF drainage or VP shunt. However, all patients subsequently deteriorated, and further progression of their disease led to the revision of the final clinical diagnosis to PSP. None of the cases had iNPH as their most likely clinical diagnosis at the time of death. One patient had a pathological diagnosis of PD, while PSP was confirmed in the other three cases. The QSBB collection is probably biased towards cases with an unclear clinical diagnosis and this could account for the high prevalence of neurodegenerative pathology among its cases diagnosed as iNPH.

All four patients had presented with gait ignition failure and freezing, a clinical picture most commonly associated with diffuse cerebrovascular disease and less commonly with PSP or PD [8, 9]. The PSP-PAGF (pure akinesia with gait freezing) subtype usually develops vertical supranuclear gaze palsy later in the disease course, and in post-mortem has milder tau pathology when compared to the classical Richardson’s syndrome [9]. In the present series, eye-movement abnormalities also occurred late in the disease course, and thus none of the patients fulfilled the NINDS-SPSP criteria for PSP in the early stages [10]. Furthermore, none of the patients had the “extended lumbar drainage protocol” which is regarded as the gold standard for the presumptive diagnosis of NPH, increasing the chance of long-term benefit from shunting to 80 % [11].

All of the patients followed a relatively benign disease course, remaining independent until a few months before death, when significant cognitive and mobility impairment, urinary incontinence and, in case 4, severe dysphagia and unintelligible speech ensued.

Excess ventricular expansion with disproportionally large ventricles for the degree of cortical atrophy can occur in PSP [12]. Recurrent falls can also lead to head injuries and subdural haematomas in patients with PSP, resulting in secondary NPH. In our series, two patients carried a risk of developing secondary NPH as a result of previous subarachnoid haemorrhage and syringomyelia warranting spinal surgery. Case 1 had sustained a left neck of femur fracture after a fall.

In a recent Mayo Clinic community-based study, iNPH was found to be relatively rare (1.19 per 100,000 per year) with an incidence approximating that of PSP [6]. Additional or alternative neurologic diagnoses later emerged in 5 of 12 patients, one having pathologically proven PSP. None of these patients had sustained definitive improvement after shunting and in the PSP case, it was speculated that the later-diagnosed disorder had been responsible for the entire clinical picture [6].

At the National Hospital for Neurology and Neurosurgery, 24 out of 87 patients (age 56–92 years) who were diagnosed as iNPH and underwent VP shunting from 2004 to 2011 showed progressive neurological features, which could not be accounted for by a diagnosis of optimally treated NPH. These 24 patients were later referred to neurologists for further assessment, and their revised clinical diagnoses are summarised in Table 2. Fourteen of the patients have since died, although none received post mortem examination. The remaining 49 patients reported benefit from shunting ranging from modest to excellent.

Table 2 Revised clinical diagnosis for patients diagnosed as iNPH (with VP shunts) in the National Hospital for Neurology and Neurosurgery
Full size table

The transient improvement following CSF removal in all cases raises the question of whether this may be a useful therapeutic strategy in PSP. Although therapeutic CSF drainage has not been used in PSP, it has been used in vascular parkinsonism (VP). Out of 40 patients with VP who were treated with large-volume CSF removal, almost 40 % reported significant gait improvement [13]. Although a placebo response cannot be excluded, one possible explanation for this initial clinical improvement could be renewal of ‘stagnant’ CSF, which increases the ability to ‘clear’ potentially toxic metabolites, such as Aβ or tau proteins. Alteration in ventricular compliance may also cause partly reversible gait dysfunction via disruption of ascending and descending periventricular tracts [7].

Cognitive impairment in NPH might also be a result of a combination of CSF dynamic disturbance and its resulting biochemical abnormalities [14]. In a recent study of 21 patients with NPH, a combination of Aβ and total tau levels in ventricular CSF predicted favourable surgical outcomes with a sensitivity of 80 % and a specificity of 82.4 % [15]. VP shunting in 35 patients with NPH led to a reduction in CSF neurofilament proteins which are markers of axonal damage, resolution of periventricular white matter hyperintensities (WMH) and clinical improvement in 28 patients 3 months after shunt insertion [16]. In another study, 12 of 17 patients with NPH had parkinsonism (71 %) which significantly improved and, in 11 patients WMH resolved following CSF tap tests or shunt operations [17]. Nevertheless, it is not known if these biochemical, radiological and clinical improvements are sustained in long-term follow-up.

Conclusion

iNPH is uncommon and a diagnosis of exclusion [2]. The absence of urinary incontinence and the presence of a supranuclear gaze palsy in a suspected case of iNPH should lead to reconsideration of the diagnosis and raise the possibility of the PAGF variant of PSP which is unresponsive to shunting.