Clinical Outcome and Striatal Dopaminergic Function After Shunt Surgery in Patients With Idiopathic Normal Pressure Hydrocephalus

Abstract

Objective

To determine changes in clinical features and striatal dopamine reuptake transporter (DAT) density after shunt surgery in patients with idiopathic normal pressure hydrocephalus (iNPH).

Methods

Participants with probable iNPH were assessed at baseline by means of clinical rating scales, brain MRI, and SPECT with [¹²³I]-N-ω-fluoropropyl-2β-carbomethoxy-3β-(4-iodophenyl)nortropane (FP-CIT). Levodopa responsiveness was also evaluated. Patients who did or did not undergo lumboperitoneal shunt were clinically followed up and repeated SPECT after 2 years.

Results

We enrolled 115 patients with iNPH. Of 102 patients without significant levodopa response and no signs of atypical parkinsonism, 92 underwent FP-CIT SPECT (58 also at follow-up) and 59 underwent surgery. We identified a disequilibrium subtype (phenotype 1) and a locomotor subtype (phenotype 2) of higher-level gait disorder. Gait impairment correlated with caudate DAT density in both phenotypes, whereas parkinsonian signs correlated with putamen and caudate DAT binding in patients with phenotype 2, who showed more severe symptoms and lower striatal DAT density. Gait and caudate DAT binding improved in both phenotypes after surgery (p < 0.01). Parkinsonism and putamen DAT density improved in shunted patients with phenotype 2 (p < 0.001). Conversely, gait, parkinsonian signs, and striatal DAT binding worsened in patients who declined surgery (p < 0.01).

Conclusions

This prospective interventional study highlights the pathophysiologic relevance of striatal dopaminergic dysfunction in the motor phenotypic expression of iNPH. Absence of levodopa responsiveness, shunt-responsive parkinsonism, and postsurgery improvement of striatal DAT density are findings that corroborate the notion of a reversible striatal dysfunction in a subset of patients with iNPH.

Motor symptoms in idiopathic normal pressure hydrocephalus (iNPH) include gait impairment, which has traditionally been considered a higher-level gait disorder (HLGD).^1-4 Parkinsonian signs involving gait, upper or lower limbs, and posture are also described.^5-15 Although parkinsonian features have shown responsiveness to shunt surgery,^6-9,14 they are not widely acknowledged within the phenotypic spectrum of iNPH. Indeed, parkinsonian signs are often seen as a red flag supporting alternative diagnoses.¹¹ Distinguishing iNPH from neurodegenerative parkinsonisms, such as Parkinson disease (PD) and atypical parkinsonisms (APs), is therefore challenging but crucial, given differences in prognosis and treatment.^16,17 It is also debated whether iNPH may predispose to or coexist with neurodegenerative disorders.^5,17-20 Some authors have argued that iNPH is an early sign of underlying neurodegenerative diseases, especially when surgery leads to short-lived benefits.^21-23

Controversy over the pathophysiology of parkinsonism in iNPH is also fueled by inconsistent evidence of reduced striatal dopamine receptors density.^9,13,24 Particularly, we have shown abnormalities of striatal dopamine reuptake transporter (DAT) binding in 62% of patients with iNPH, as detected by SPECT with [¹²³I]-N-ω-fluoropropyl-2β-carbomethoxy-3β-(4-iodophenyl)nortropane (FP-CIT).¹⁵ It is unclear whether this striatal involvement reflects a concomitant neurodegenerative process or direct mechanical effects induced by altered CSF dynamics.^9,13,15,24 Assessment of changes in striatal dopamine receptors after shunt treatment might help to clarify the pathophysiology of striatal involvement in iNPH. However, longitudinal molecular imaging studies are lacking.

We report clinical and FP-CIT SPECT findings of a prospective longitudinal cohort study of patients with iNPH, comparing patients who did vs those who did not undergo shunt surgery.

Methods

Clinical and MRI Assessment

Between January 2016 and March 2018, we consecutively evaluated 115 patients with probable iNPH, all diagnosed according to the iNPH International Guidelines,¹ at the Parkinson's Disease and Movement Disorders Unit of the Mondino Foundation in Pavia (Italy).

At baseline, we assessed gait, cognition, balance, and continence using the iNPH Rating Scale (iNPHRS), where lower values correspond to more severe dysfunction.²⁵ In keeping with the literature,^4,15,26,27 we identified 2 gait patterns: a disequilibrium subtype of HLGD, that is, a wide base and externally rotated feet, and a locomotor subtype of HLGD, that is, start hesitation, freezing of gait, shuffling steps, and en bloc turning. We assessed parkinsonian signs using the motor section of the Movement Disorders Society–Unified Parkinson's Disease Rating Scale (MDS-UPDRS III).²⁸ We considered the following scores, obtained by summing single items of the MDS-UPDRS III: axial score (item 3.1, neck score of item 3.3, and items 3.9–3.13), limb score (limb scores of item 3.3, items 3.4–3.8, and 3.15–3.18), tremor score (items 3.15–3.18), rigidity score (item 3.3), bradykinesia score (items 3.4–3.8 and 3.14), upper body score (upper limb scores of item 3.3, items 3.4–3.6, 3.15, 3.16, and upper limb scores of item 3.17), and lower body score (lower limb scores of item 3.3, items 3.7, 3.8, and lower limb scores of item 3.17).^10,29 We defined bradykinesia as slowness with decrement of amplitude or speed during repetitive movements and parkinsonism as bradykinesia combined with rest tremor, rigidity, or both.³⁰

All patients underwent brain MRI (3T Skyra; Siemens) at baseline. We calculated the Evans index and looked for the following supportive features of iNPH^1,2: acute callosal angle at the level of the posterior commissure; aqueductal or fourth ventricular flow void; disproportionately enlarged subarachnoid space hydrocephalus (DESH), that is, the obliteration of the high-convexity sulci; and dilation of the Sylvian fissures.³¹ We also assessed periventricular and deep white matter hyperintensities using the Fazekas scale.³²

All the patients were treated for 3 months with levodopa 600 mg daily (200 mg 3 times daily) and were assessed monthly using the MDS-UPDRS III. At the end of this 3-month period, all the patients underwent an acute challenge test with levodopa 250 mg after an overnight drug withdrawal. Both chronic and acute levodopa challenge tests were performed to increase the sensitivity of the factor “levodopa responsiveness.” A significant improvement with levodopa was defined as a greater than 30% reduction in the MDS-UPDRS III total score.³³ We also evaluated the presence of additional red flag signs suggestive of APs. Patients with a positive levodopa challenge test indicative of PD or with signs supporting APs were excluded from further analyses. The remaining 102 patients discontinued levodopa therapy and underwent a tap test (i.e., large-volume lumbar puncture). A positive tap test response was defined as an at least 10% improvement of 10-meter seconds or steps.³¹ All 34 patients with a negative response to the tap test underwent an external lumbar drainage, and all showed a positive response, which was defined as for the tap test.³¹

We recommended shunt treatment to all 102 patients, but 43 declined surgery for logistical reasons (far from our center and not willing to undergo valve programming) or fear of potential adverse events related to the procedure. Fifty-nine patients underwent the placement of a lumboperitoneal shunt with programmable valve (Medtronic PS Medical Strata NSC LPS) on average 4 months after the baseline assessment. All surgical procedures were performed at the Neurosurgery Unit of the San Matteo Foundation in Pavia. No shunt failures, infections, or revisions occurred.³¹

All 102 patients were then followed up for 2 years through scheduled annual clinical evaluations.

Molecular Imaging Acquisition and Analysis

At baseline, 92 patients without evidence of PD or APs underwent FP-CIT SPECT. Among these patients, 58 also repeated FP-CIT SPECT at the 2-year follow-up. Scans were acquired 180 minutes after injection of 182.4 ± 3.6 MBq of FP-CIT on a dual-headed SPECT system (Infinia; GE Healthcare) equipped with a fan-beam collimator (step-and-shoot mode, 1 frame/4°/45 seconds, photopeak window of 159 keV ±15%, matrix 128 × 128). Reconstructions were performed on a Xeleris platform (GE Healthcare) with filtered back projection prefiltering with Butterworth cutoff of 0.55 cycles/cm and order 10. SPECT data were processed and analyzed using the relevant toolboxes of PMOD image analysis software version 4.0 (PMOD Technologies Ltd.). Semiquantitative evaluation of SPECT data was employed to obtain the specific binding ratio (SBR) of FP-CIT in striatal volumes of interest (VOIs). SPECT data were normalized to a dedicated template in the standard anatomical space of the Montreal Neurologic Institute. Based on the Automated Anatomical Labeling Atlas, paired VOIs for the putamen and caudate nucleus and large bilateral VOIs for the occipital lobe were obtained. Then, VOI-based partial volume correction was performed. Finally, we calculated the average regional binding values in the VOIs and the SBR for the putamen and caudate nucleus using the occipital cortex as reference region.^15,34

Statistical Analysis

Data were evaluated for normal distribution and equality of variances using the Shapiro-Wilk test and Levene test, respectively. Motor phenotypes were compared using Pearson χ² test, a t test, or the Wilcoxon rank-sum test, as appropriate. The Bonferroni correction for multiple comparisons was also applied. Pearson correlation coefficient was used for bivariate analyses. Within-phenotype clinical differences at follow-up were verified by means of linear mixed-effects models with repeated measures. Data obtained from the dependent variables (i.e., single items and total scores of the iNPHRS and MDS-UPDRS III) were log-transformed for normality. Treatment (2 levels: shunt and no shunt), time (3 levels: baseline, 1-year follow-up, and 2-year follow-up), and the interaction between treatment and time were modeled as fixed effects, whereas patients were considered as a random factor. Final models were fitted using the restricted maximum likelihood method. If significant differences of the factor “time” were identified, a paired sample t test with the Bonferroni correction was applied. Phenotype conversion at follow-up was evaluated using the McNemar test. Within-phenotype differences in DAT binding at follow-up were assessed with a paired sample t test or the Wilcoxon signed-rank test, as appropriate. Mean percentage rates of improvement after shunt were calculated as follows: [(b − a)/a]  × 100, where a is the mean baseline score and b is the mean follow-up score of a clinical rating scale. Values of p < 0.05 were considered significant. Statistical analyses were performed using JMP Pro 14.0 (SAS Institute Inc.).

Standard Protocol Approvals, Registrations, and Patient Consents

The Ethics Committee of the Casimiro Mondino Foundation approved the study. All patients gave written informed consent to all study procedures and to personal data processing for research purposes, according to the Declaration of Helsinki.

Data Availability

Anonymized data will be shared upon appropriate request from any qualified investigator.

Results

Baseline Features of iNPH

Figure 1 is a flow chart summarizing the clinical features of the whole sample of patients with iNPH at baseline. In particular, 102 (88.7%) patients showed neither a significant improvement with levodopa (mean ± SD and range of the MDS-UPDRS III total score change: 3.2 ± 2.6%, 0%–6.6% after 3-month trial; 3.4 ± 2.9%, 0%–6.8% after acute challenge test) nor any red flag signs suggestive of APs. Among these patients without evidence of PD or APs, we identified 2 motor phenotypes of iNPH: a disequilibrium subtype of HLGD (phenotype 1: 34 patients, 33.3%) and a locomotor subtype of HLGD, including freezing of gait (phenotype 2: 68 patients, 66.7%).

Figure 1. Clinical Features of the Whole Sample of Patients With Idiopathic Normal Pressure Hydrocephalus (iNPH).

Parkinson disease (PD) and progressive supranuclear palsy (PSP) were diagnosed according to the Movement Disorders Society criteria (designed in 2015 and 2017, respectively), dementia with Lewy bodies (DLB) according to the criteria of the 4th consensus report of the DLB Consortium (2017), and corticobasal syndrome (CBS) according to the criteria of Armstrong and colleagues (2013). In all patients with PD, the motor improvement after 3-month levodopa trial was comparable to what we observed after acute levodopa challenge test. AP = atypical parkinsonism; HLGD = higher-level gait disorder.

Table 1 lists the demographic, clinical, and neuroimaging findings of the iNPH motor phenotypes at baseline. They showed no differences in age at onset of symptoms, disease duration, or comorbidities. Patients with phenotype 2 showed more impaired gait and balance and more severe and more frequent cognitive and urinary symptoms as compared to patients with phenotype 1. The predominance of parkinsonian signs in patients with phenotype 2 reflected differences detected on the MDS-UPDRS III. Indeed, patients with phenotype 2 had higher scores for axial and limb impairment, tremor, rigidity, bradykinesia, and upper and lower body involvement than patients with phenotype 1 (table 1).

Table 1.

Demographic, Clinical, and Neuroimaging Features of Idiopathic Normal Pressure Hydrocephalus Motor Phenotypes at Baseline

At baseline, 32 patients with phenotype 1 and 60 with phenotype 2 underwent FP-CIT SPECT. DAT density of the putamen and caudate nucleus was lower in patients with phenotype 2 than in those with phenotype 1 (table 1). SBR values of the caudate nucleus correlated with the iNPHRS gait score in both motor phenotypes (figure 2, A and B). By contrast, caudate DAT density did not correlate with iNPHRS scores for cognition, balance, or continence. We did not find any correlation between putamen SBR values and iNPHRS scores. Furthermore, putamen and caudate DAT density correlated with the MDS-UPDRS III total score only in patients with phenotype 2 (figure 2C). In this subgroup, putamen and caudate SBR values also correlated with MDS-UPDRS III scores for the upper body and lower body, respectively (figure 2, D and E).

Figure 2. Scatter Plots of Striatal Specific Binding Ratio (SBR) Values and Clinical Scores.

Pearson correlation coefficients with regression lines (solid lines) and 95% confidence interval (dashed lines) for the iNPH Rating Scale (iNPHRS) gait score in phenotype 1 (A) and phenotype 2 (B), and for total score (C), upper body score (D), and lower body score (E) of the Movement Disorders Society–Unified Parkinson's Disease Rating Scale, motor section (MDS-UPDRS III) in phenotype 2. Phenotype 1 is the disequilibrium subtype and phenotype 2 is the locomotor subtype of higher-level gait disorder.

Striatal DAT density did not correlate with baseline MRI features of ventriculomegaly (the Evans index, acute callosal angle, DESH) or white matter involvement (Fazekas scale scores). The motor phenotypes did not differ with respect to these MRI findings (table 1).

Follow-up of Patients With iNPH With Disequilibrium Subtype of HLGD

The baseline and follow-up scores of the 34 patients with phenotype 1 are shown in table 2. Patients who did vs those who did not undergo shunt placement did not differ with regard to baseline scores of the iNPHRS and MDS-UPDRS III. Seventeen of them (50%) underwent surgery. After 1 year they showed improvements in gait, balance, cognition, continence, axial and limb impairment, bradykinesia, and upper and lower body involvement scores. The effectiveness of the surgery, measured as improvements on baseline values, persisted at 2 years, without any difference with respect to the 1-year evaluation. None of them converted to phenotype 2.

Table 2.

Baseline and Follow-up Scores of Patients With Idiopathic Normal Pressure Hydrocephalus With Disequilibrium Subtype of Higher-Level Gait Disorder (Phenotype 1)

Twelve of the shunted patients repeated FP-CIT SPECT at the 2-year assessment. Compared with the baseline data, they showed increased caudate SBR values (right: p = 0.006; left: p = 0.002), but no differences in the putamen values (figure 3, A and B).

Figure 3. FP-CIT SPECT Differences at Follow-up Compared With the Baseline Evaluation in Phenotype 1.

Comparison of striatal specific binding ratio (SBR) values between T0 and T2 in patients who did (A) or did not (C) undergo shunt. Box plots depict median and interquartile range; the ends of the whiskers stand for minimum and maximum values. Horizontal bars indicate significant differences (*p < 0.01, **p < 0.001). Representative FP-CIT SPECT scans, at T0 and T2, of a patient who underwent surgery (B) and of a patient who declined shunt treatment (D) are shown. Phenotype 1 is the disequilibrium subtype of higher-level gait disorder. T0 = baseline; T2 = 2-year follow-up.

Conversely, at the 1-year evaluation, the 17 patients with phenotype 1 who had declined shunt treatment showed worse scores for gait, balance, axial and limb impairment, bradykinesia, and upper and lower body involvement. This same deterioration persisted at the 2-year follow-up as compared with both the baseline and the 1-year evaluation. Rigidity was also worse at the 2-year follow-up with respect to the baseline assessment. These patients showed no changes in cognition or continence. Of the 17 patients with phenotype 1 not treated with surgery, 4 (23.5%) subsequently converted to phenotype 2 (one patient after 1 year, 3 after 2 years). This phenotype conversion was significant (p = 0.008).

Thirteen of the nonshunted patients with phenotype 1 repeated FP-CIT SPECT after 2 years. With respect to the baseline evaluation, they showed decreased caudate DAT density (right: p = 0.002; left: p < 0.001), but no differences with regard to the putamen (figure 3, C and D).

None of the patients with phenotype 1 developed PD or APs during the 2-year follow-up.

Follow-up of Patients With iNPH With Locomotor Subtype of HLGD

The baseline and follow-up scores of the 68 patients with phenotype 2 are listed in table 3. Patients who did vs those who did not undergo surgery did not differ with respect to baseline scores of the iNPHRS and MDS-UPDRS III. Forty-two of them (61.8%, p > 0.05 vs phenotype 1) underwent shunt placement. At the 1-year evaluation, these patients showed improvements in gait, balance, axial and limb impairment, tremor, rigidity, bradykinesia, and upper and lower body involvement scores. These improvements on the baseline values persisted at 2 years, without showing any difference in comparison with the 1-year assessment. Cognition and continence did not change. By the 1-year follow-up, 15 (35.7%) of the shunted patients with phenotype 2 had converted to phenotype 1. This phenotype conversion was significant (p < 0.001).

Table 3.

Baseline and Follow-up Scores of Patients With Idiopathic Normal Pressure Hydrocephalus With Locomotor Subtype of Higher-Level Gait Disorder (Phenotype 2)

Twenty-one of the shunted patients repeated FP-CIT SPECT at the 2-year evaluation. They showed higher SBR values of both the putamen (right: p < 0.001; left: p < 0.001) and the caudate nucleus (right: p = 0.003; left: p = 0.006) compared with the baseline findings (figure 4, A and B).

Figure 4. FP-CIT SPECT Differences at Follow-up Compared With the Baseline Evaluation in Phenotype 2.

Comparison of striatal specific binding ratio (SBR) values between T0 and T2 in patients who did (A) or did not (C) undergo surgery. Box plots depict median and interquartile range; the ends of the whiskers stand for minimum and maximum values. Horizontal bars indicate significant differences (*p < 0.01, **p < 0.001). Representative FP-CIT SPECT scans, at T0 and T2, of a patient who underwent shunt treatment (B) and of a patient who declined surgery (D) are shown. Phenotype 2 is the locomotor subtype of higher-level gait disorder. T0 = baseline; T2 = 2-year follow-up.

On the other hand, the 26 patients with phenotype 2 who declined surgery showed a worsening of gait, balance, cognition, continence, axial and limb impairment, tremor, rigidity, bradykinesia, and upper and lower body involvement after 1 year. At 2 years, the deterioration had worsened further as compared with both the baseline and the 1-year evaluation. None of them converted to phenotype 1.

Twelve of the nonshunted patients with phenotype 2 repeated FP-CIT SPECT after 2 years. In comparison with the baseline values, they showed lower DAT density of both the putamen (right: p = 0.004; left: p = 0.005) and the caudate nucleus (right: p = 0.004; left: p = 0.003) (figure 4, C and D).

None of the patients with phenotype 2 developed PD or APs during the 2-year follow-up.

Clinical Follow-up After Surgery: Comparison Between Motor Phenotypes

The baseline and follow-up scores of the shunted patients with phenotypes 1 and 2 are shown in tables 2 and table 3, respectively. Among the shunted patients, the mean rates of clinical improvement were higher in those with phenotype 2 vs phenotype 1 with regard to gait (99.8% vs 65.4%, p = 0.014 at 1 year; 102.2% vs 71.6%, p = 0.016 at 2 years), upper body impairment (88.1% vs 48.8%, p = 0.003 at 1 year; 88.3% vs 49.3%, p = 0.004 at 2 years), and lower body involvement (71.1% vs 45.4%, p = 0.011 at 1 year; 71.8% vs 46.4%, p = 0.012 at 2 years).

Conversely, the phenotype 2 and phenotype 1 patients showed similar rates of improvement in the iNPHRS balance score (57.3% vs 40.4%, p = 0.134 at 1 year; 58.2% vs 40.6%, p = 0.124 at 2 years) and MDS-UPDRS III axial score (38.1% vs 31.0%, p = 0.331 at 1 year; 38.9% vs 32.5%, p = 0.386 at 2 years).

Discussion

This prospective study highlights the relevance of striatal dopaminergic function in the pathophysiology of motor disturbances in iNPH. We showed that striatal DAT density correlates with gait impairment and parkinsonian signs, and that it improves after shunt, whereas it worsens in patients who decline surgery. Of note, shunt responsiveness supports the argument that parkinsonian features should be considered part of the phenotypic spectrum of iNPH, once mimic neurodegenerative parkinsonisms have been excluded. The improved striatal dopaminergic activity observed after surgery confirms a reversible dysfunction of presynaptic dopamine pathways rather than an underlying neurodegeneration, in keeping with the evidence of restored immunoreactivity of nigral neurons in hydrocephalic rats after shunt.³⁵ In this context, a reversal of nigrostriatal dopaminergic dysfunction might reflect an improvement of stretch/compression injury caused by CSF pressure against the ventricular walls.³⁶ Indeed, the proximity of the caudate nucleus to the dilated ventricles may explain why this is the portion of the striatum most frequently involved, in agreement with an MRI study showing a smaller caudate volume in patients with iNPH compared with healthy controls.³⁷ In addition to basal ganglia, a volume reduction has been observed in other subcortical structures, such as thalamus and hippocampus, and it was partially reversed by shunt treatment.³⁸ Alternatively, the recovery of striatal dopaminergic function could be due to enhanced clearance of potentially toxic metabolites after surgery,³⁹ although this mechanism is highly debated. Consistently with these hypotheses, further progression of these mechanical and metabolic alterations could be an explanation for the worsening of the striatal presynaptic dopaminergic activity as described in patients who declined surgery.

Moreover, involvement of the nigrostriatal dopaminergic pathway may not be the only pathophysiologic mechanism underlying motor disturbances in iNPH. Indeed, the findings of levodopa-unresponsive parkinsonism in iNPH may conceivably point to a concomitant alteration of striatal postsynaptic dopaminergic activity.⁵ This idea is supported by PET studies reporting reduced density of postsynaptic D2 receptors in the putamen²⁴ and their enhanced expression after shunting in iNPH.⁴⁰

In this study, we also characterized the long-term evolution of the 2 motor phenotypes of iNPH. These phenotypes are most likely expressions of the same pathophysiologic continuum, as suggested by the postshunt conversion of some patients from the locomotor (phenotype 2) to the disequilibrium (phenotype 1) subtype of HLGD, and by the progression to phenotype 2 of some nonshunted patients initially displaying phenotype 1. These observations support the hypothesis that phenotype 2 is a more severe form of iNPH, an idea further corroborated by the different cognition and continence outcomes between the 2 motor phenotypes. Nevertheless, both phenotypes showed improvements of gait and parkinsonian signs after surgery, which were actually greater among patients with phenotype 2. In keeping with the notion that gait problems are more shunt-responsive,⁴¹ the balance response to surgery was less pronounced in both subgroups of patients. The lack of significant cognitive improvement after shunt treatment in both motor phenotypes corroborates our previous observations,³¹ although this matter remains debated.

We identified appendicular parkinsonism in the locomotor subtype of HLGD, specifically in 66.7% of our patients with iNPH without evidence of PD or APs. This prevalence is in line with the rate reported in several studies, which ranges from 60% to 71%.^6,7,9,12 The motor phenomenology observed in our sample encompasses a progressive decrement in amplitude and speed during repetitive movements and therefore closely resembles bradykinesia of a “true” parkinsonism.³⁰ This notion differs from the “pseudo-parkinsonism” previously described in iNPH and defined as reduced amplitude and speed of movements in absence of “sequence effect.”⁴² As a result, the differential diagnosis can be challenging, in particular when iNPH is suspected in patients with PD or APs, as found in 13 (11.3%) patients in our sample. Nonetheless, parkinsonism in iNPH does not significantly improve with levodopa, as reported by several authors recommending a levodopa challenge test for the differential diagnosis with PD.^6,16 The presence of parkinsonian signs should not preclude surgery in patients with iNPH, who can potentially show substantial and prolonged improvement after shunt placement, as described by other reported cases of shunt-responsive parkinsonism in iNPH.^7-9 However, previous studies evaluated small samples of patients over short follow-up periods. In our patients, the sustained efficacy of shunt treatment and the absence of signs supporting alternative diagnoses at the 2-year follow-up are noteworthy findings given the many controversies over the diagnosis of iNPH.^21-23

The clinical differences between the 2 motor phenotypes were substantiated by SPECT findings showing a more severe striatal presynaptic dopaminergic loss in patients with phenotype 2. Indeed, in this cohort the putaminal involvement could have determined the presence and severity of upper body parkinsonism, whereas the caudate dysfunction may have contributed to more severe gait disturbances, including freezing of gait.

Although a few authors have explored striatal DAT binding in patients with iNPH with parkinsonism, finding it to be reduced in a variable proportion,^9,13 only our recent study has investigated this topic in depth, showing a symmetric and prominent DAT deficit in the caudate nucleus, and an association between striatal DAT binding and severity of parkinsonian signs in iNPH.¹⁵ What is well known, on the other hand, is the inverse correlation of striatal DAT density with bradykinesia, rigidity, gait, and posture in PD, without significant differences between the putamen and caudate nucleus.⁴³ Our findings further highlight the role of striatal dopamine⁴⁴ and particularly of the caudate nucleus⁴⁵ in gait control. Of relevance, walking speed in elderly patients was positively associated with volume of the caudate nucleus.⁴⁶ Instead, in patients with PD, the presence and severity of freezing of gait were related to a prominent dopaminergic denervation⁴⁷ and lower volume⁴⁸ of the caudate nucleus. Conversely, 2 PET studies linked postsynaptic dopaminergic dysfunction of the putamen, but not of the caudate nucleus, to gait impairment in iNPH along with preserved presynaptic activity of the nigrostriatal dopaminergic system.^24,40 However, these discrepancies with our study may be due to differences in sample size and in the choice of gait scoring method as well as the presynaptic radioligand used.

Our study has several limitations. First, the trial was neither blinded nor randomized. This is particularly relevant with respect to the comparisons with patients who declined surgery. Although their baseline characteristics did not differ from those of patients undergoing shunting, a possible selection bias cannot be ruled out. Interestingly, disease severity did not play a role in patients' decision to undergo surgery (e.g., a similar proportion of phenotype 1 and 2 opted for shunting), in keeping with the observation that none of the patients initially declining surgery reconsidered this option during the follow-up period despite a progression of their symptoms. Second, a longer follow-up would have been desirable to rule out a “dual” pathology, since a possible hydrocephalic presentation of APs has been described.^21-23 Indeed, several authors suggest that a superimposed neurodegenerative process might explain the mechanism of short-lived benefits after surgery in a subset of patients with iNPH.^21-23 Although we did not observe any decline in shunt benefit, an underlying neurodegenerative process cannot be entirely excluded in our patients given the relatively short follow-up and the lack of pathology data. Third, the lack of a SPECT-CT prevented a proper coregistration of SPECT and MRI. Therefore, VOIs positioning could have been affected by the anatomical distortion due to ventriculomegaly. Still, we can reasonably rule out this bias, since the Evans index did not significantly change over time in patients who did and did not undergo shunt surgery.³¹ Fourth, SPECT has lower resolution and worse quantitative capacity compared with PET, but is widely used in both clinical practice and research projects. Fifth, downregulation of striatal DAT binding in iNPH can result in underestimation of the true density of nerve terminals.⁴⁹ Alternative radioligands, such as [¹⁸F]-dihydroxyphenylalanine (DOPA), could be useful given their closer correlation with the functional integrity of dopaminergic terminals.⁵⁰ Finally, the cognitive evaluation was limited to the related iNPHRS item, whose score derives from 3 reliable and valid neuropsychological tests (i.e., the Grooved Pegboard, the Rey Auditory Verbal Learning Test, and the Stroop test).²⁵ A more extensive cognitive assessment would therefore be recommended to verify longitudinal changes and correlations with neuroimaging findings.

This prospective study provides clinical and imaging insights into the pathophysiology of iNPH. The locomotor subtype emerged as the most severe and frequent motor expression of iNPH. Improvement of nigrostriatal dopaminergic function may be an important pathophysiologic mechanism underlying clinical responsiveness to shunt treatment. Postsurgery improvement of striatal DAT density supports a reversible striatal dysfunction. Of note, the putamen and caudate nucleus could contribute differently to motor disturbances in iNPH. Studies with multiple radioligands may help to identify additional pathways involved in iNPH, deepening our understanding of the pathophysiology of this complex disease.

Glossary

AP

atypical parkinsonism

DAT

dopamine reuptake transporter

DESH

disproportionately enlarged subarachnoid space hydrocephalus

FP-CIT

[¹²³I]-N-ω-fluoropropyl-2β-carbomethoxy-3β-(4-iodophenyl)nortropane

HLGD

higher-level gait disorder

iNPH

idiopathic normal pressure hydrocephalus

iNPHRS

iNPH Rating Scale

MDS-UPDRS III

Movement Disorders Society–Unified Parkinson's Disease Rating Scale, motor section

PD

Parkinson disease

SBR

specific binding ratio

VOI

volume of interest

Appendix. Authors

Study Funding

This study was approved and funded by Italian Ministry of Health and co-funded by the Health Service of Lombardy (project code number RF-2013-02355908: “Idiopathic normal pressure hydrocephalus [iNPH], parkinsonism and dementia: improving the accuracy of diagnosis and the patient care to reverse the symptomatology: neurodegeneration, phenotypes and outcome measures”).

Disclosure

J. Brumberg received a scholarship from Deutsche Forschungsgemeinschaft. R. Ceravolo is part of advisory boards for AbbVie, UCB, and Zambon and received honoraria from AbbVie, General Electric, Lundbeck, Lusofarmaco, UCB, and Zambon. I.U. Isaias received honoraria from Medtronic, as well as grants from Medtronic, Newronika, Deutsche Forschungsgemeinschaft, Michael J. Fox Foundation for Parkinson's Disease, Interdisciplinary Research Centre for Clinical Research Würzburg, Bavaria California Technology Center, Grigioni Foundation for Parkinson's Disease, and Fondazione Europea Ricerca Biomedica. A. Fasano provided consultancies to Abbott, AbbVie, Boston Scientific, Chiesi Farmaceutici, Ipsen, Medtronic, Sunovion, and UCB; is part of advisory boards for Abbott, AbbVie, Boston Scientific, and Ipsen; received honoraria from Abbott, AbbVie, Boston Scientific, Chiesi Farmaceutici, Ipsen, Medtronic, Sunovion, and UCB; and received grants from AbbVie, Boston Scientific, Medtronic, University of Toronto, and Weston Foundation. C. Pacchetti provided consultancies to AbbVie, Boston Scientific, and Medtronic and received honoraria from AbbVie, Boston Scientific, Lusofarmaco, Medtronic, and Zambon. The other authors declare no competing interests. Go to Neurology.org/N for full disclosures.