Exploring the Association Between Visual Field Testing and CERAD Neuropsychological Battery in Idiopathic Normal Pressure Hydrocephalus Patients

Abstract

Background:

Association between visual field test indices and The Consortium to Establish a Registry for Alzheimer’s Disease Neuropsychological Battery (CERAD-NB) is unknown. Idiopathic normal pressure hydrocephalus (iNPH) patients provide a unique set of patient data for analysis.

Objective:

To assess the reliability of visual field testing using the CERAD-NB in patients with iNPH and to investigate the association between visual field test results and cognitive function.

Methods:

62 probable iNPH patients were subjected to comprehensive ophthalmological examination, ophthalmological optical coherence tomography imaging studies, visual field testing, and CERAD-NB. Based on visual field indices, the patients were divided into two groups: unreliable (n = 19) and reliable (n = 43). Independent T-test analysis was performed to examine the relationship between visual field test results and cognitive function. Pearson Chi-square test was used for non-continuous variables.

Results:

The unreliable group performed worse in CERAD-NB subtests compared to the reliable group. Statistically significant differences were observed in nine out of ten subtests, with only Clock Drawing showing no statistical significance. Pairwise comparison of the groups showed no statistical significance between amyloid-β (Aβ) biopsy, hyperphosphorylated tau biopsy, apolipoprotein E allele or the ophthalmological status of the patient. But there was a statistically significant difference in cerebrospinal fluid Aβ₄₂ and age between the groups.

Conclusions:

Patients with unreliable visual field tests performed worse on CERAD-NB subtests. CERAD-NB subtests do not provide a specific cut-off value to refrain patients from visual field testing. Should patients with unreliable visual field tests be screened for cognitive impairment?

Keywords

Alzheimer’s disease amyloid-β CERAD-NB cognitive impairment normal pressure hydrocephalus optical coherence tomography visual field tests

INTRODUCTION

Idiopathic normal pressure hydrocephalus (iNPH) is a chronic disease among the elderly. It is characterized by the classic triad (also known as Hakim’s triad) of symptoms consisting of gait disturbance, cognitive impairment, and urinary incontinence. It is also associated with radiologically verified ventriculomegaly [1]. Despite the severity of the symptoms, all of them can improve following shunt surgery [2]. Several cerebrospinal fluid (CSF) hydrodynamic mechanisms have been proposed to contribute to the development of iNPH. Yet, the exact cause remains elusive [3]. The prevalence of iNPH has been estimated to be 0.2% in the age group from 70–79 and 5.9% of 80 years and older [4].

In the early stages of iNPH, cognitive symptoms significantly impair memory and psychomotor speed [5], cause inattention, and lead to neuropsychiatric symptoms such as lack of spontaneity and progressive apathy [6]. The patients with iNPH also performed worse on neuropsychological tests such as CERAD compared to the general healthy population [7]. In addition, occurrence of Alzheimer’s disease (AD) as well as neuropathological findings related to AD are increased among iNPH patients [8]. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) was originally established in 1986 by the funding of the National Institute on Aging to develop standardized and validated measures for the assessment of AD [9]. CERAD-neuropsychological battery (NB) was established to recognize early cognitive impairment related to AD. In Finland CERAD-NB was introduced in 1999. After introduction it has been used widely in primary health care as a screening instrument to detect dementia and milder cognitive problems as the Mini-Mental State Examination (MMSE) lacks delayed memory task, which is found to be crucial in detecting early AD [10].

Recently it was demonstrated that patients with iNPH exhibit a notable reduction in ganglion cells [11]. Decrease of ganglion cells is associated with a variety of neurological and ophthalmological conditions, including normal tension glaucoma (NTG). NTG is characterized by progressive optic nerve damage and visual field defects despite intraocular pressure (IOP) being within the normal range (<21 mmHg). The pathophysiology of NTG is not fully understood, but various theories have been proposed, including the role of pressure gradient across the lamina cribrosa, which is located between the intraocular and intracranial compartments [12]. Earlier studies have noted that the rate of NTG is remarkably higher in iNPH patients and patients with both NTG and iNPH expressed shallow optic disc cupping compared to NTG patients without iNPH [13]. There is also a possible link between the shunt treatment of iNPH and the development of NTG [14].

Standard automated perimetry (SAP) is a common method used in clinical practice to assess the visual field. SAP manufacturers usually provide cut-off values to deem a test unreliable, but in literature such values may vary [15]. Humphrey Field Analyzer (HFA, Zeiss Humphrey Systems) with Swedish interactive threshold algorithm (SITA) uses a cutoff value of 15% for false positives (FP) and 20% for fixation loss (FL). False negative (FN) responses are not considered flagging results as unreliable due to multiple studies finding elevated FN rates in pathological visual fields even with attentive patients. FLs are defined as patients reacting to a stimulus presented in the physiological blind spot which they should not be able to see when fixating on the central target. FPs are flagged when patients react to a nonexistent stimulus. FNs are considered to occur when patients do not react to suprathreshold stimulus presented in previously recognized areas. FNs, especially in healthier eyes can also suggest that poor attention was given during the test.

In the most recent version of the Humphrey Field Analyzer primer, it is now recommended to evaluate the reliability indices as a continuum and indices will be adjusted in newer software versions [16]. Recently studies have explored the degree to which FL, FP and FN contribute to reliability by quantitatively calculating their impacts on mean deviation (MD). MD calculates deviation from patient’s age-matched results across all test locations. Patients who require brighter stimuli for a reaction will have negative MD scores and patients who can detect dimmer stimuli will have positive MD scores. Pattern standard deviation (PSD) is used to depict focal defects. Higher PSD values are observed in visual fields with focal field defects and normal or diffusely abnormal visual fields will have low PSD. Visual Field Index (VFI) is a numerical metric that summarizes the overall health of a patient’s visual field in a single value ranging from 0–100%. VFI is calculated by considering the sensitivity values across the whole tested visual field and providing a summary of the visual field test results. A VFI score of 0% indicates a completely blind field in the tested area, while a score of 100% indicates a normal visual field. In relatively healthy patients FL did not meaningfully impact reliability. Every FP decreases reliability, but especially after >20% of catch trials. FN have less impact on reliability than FP but should be considered significant after >25%. Also, an increase in test duration beyond >2–3 of typical time taken could be considered as an indicator of poor reliability [17].

A relationship between visual field reliability indices and cognitive impairment has been shown in glaucoma patients by using the Clock Drawing test [18]. Visual field testing is a crucial tool used in ophthalmology to detect and monitor pathological processes affecting the optic nerve and visual pathways, such as glaucoma and retinal disorders. The aforementioned conditions are frequently observed in elderly patients who are more susceptible to cognitive impairments.

Optical coherence tomography (OCT) is a non-invasive imaging technique that is extensively utilized in ophthalmology for the diagnosis, treatment guidance, and disease monitoring of various ocular diseases such as glaucoma, age-related macular degeneration, and diabetic retinopathy. OCT enables in vivo cross-sectional imaging of the retina. Technological advancements in newer spectral-domain OCT (SD-OCT) allow high-resolution 3D imaging for assessing the optic nerve head (ONH), retinal nerve fiber layer (RNFL), and macula. Reduced RNFL thickness has been established in many neurodegenerative disorders [19].

The aim of this study was to evaluate the reliability of visual field testing in iNPH in relation to cognitive status.

MATERIALS AND METHODS

This prospective cross-sectional study was conducted at Kuopio University Hospital. The study was approved by The Kuopio University Hospital Research Ethical Committee and was conducted according to the tenets of Declaration of Helsinki. We obtained written informed consent from all subjects in this study. Our objective was to evaluate the success of visual field tests and other ophthalmological findings in patients with iNPH and compare it to the cognitive profile of the patients using the CERAD-NB. Patients were divided into two groups based on their visual field test results for the comparison. We examined 62 Finnish patients (Fig. 1), both male and female, aged between 60 and 90 years, who had a history of probable iNPH and existing brain biopsy AD-related pathology evaluation, along with apolipoprotein E (APOE) genotype and CSF AD biomarkers when available. The patients were selected from Kuopio University Hospital NPH-register and all of them had been treated with a shunt at Kuopio University Hospital Department of Neurosurgery between 2/2012 and 12/2018. All patients in this study had a Mini-Mental State Examination (MMSE) score of 20 or higher at their latest MMSE result and a Clinical Dementia Rating (CDR) score of 1 or lower [20]. All ophthalmic and cognitive examinations were performed between 2/2018 and 2/2019. The examinations for patients were conducted at intervals of no more than four months, and all the patients had undergone shunting procedures prior to these examinations. Exclusion criteria for this study consisted of compromised well-being of the patients, different contraindications for lumbar puncture as well as serology positive for human immunodeficiency virus and hepatitis B or C.

Fig. 1

Patient selection flowchart. iNPH, idiopathic normal pressure hydrocephalus; MMSE, Mini-Mental State Examination; visual field unreliable, false positive rate ≥15% and/or false negative rate ≥25% in either eye; visual field reliable, false positive rate <15% and false negative rate <25% in both eyes.

iNPH diagnostics

The international diagnostic guidelines for iNPH classify the diagnosis into three distinct categories: probable, possible, and unlikely. Patients categorized as probable iNPH are diagnosed based on clinical history, brain imaging, physical findings, and physiological criteria [21]. All the patients participating in this study fulfilled the criteria for probable iNPH.

Comorbidities

To conduct a comparative analysis of significant co-morbidities, we collected data on the utilization of hypertension and diabetes medications at the time of patient enrollment in the NPH registry. Hypertension medication was defined as the use of beta-blockers, angiotensin-converting enzyme inhibitors, or angiotensin receptor blockers. Diabetes medication criteria were met if patients received any oral medication or insulin for type 2diabetes.

Cognitive examination

All the patients underwent CERAD-NB examination. In the Finnish version of the test ten subtests are included. Verbal Fluency (animals in 60 s), The 15-Item Boston Naming Test (scale of 0–15), MMSE (scale of 0–30), Constructional Praxis (scale of 0–11), Delayed Constructional Praxis (scale of 0–11), Word List Learning (scale of 0–30), Word List Recall (scale of 0–10), Word list Recall Savings (%), Word List Recognition (%, scale of 0–100) and Clock Drawing (scale of 0–6). All the CERAD-NB examinations were performed by an experienced nurse. In accordance with the Finnish educational system, where basic education comprises 9 years of studies, patients were dichotomized into two groups based on their level of education: those with 9 years or less, and those with more than 9 years of education.

Biopsy procedure and immunohistochemistry

During shunt surgery brain biopsies were taken and analyzed using the authorized protocol. The biopsies were taken from the right frontal cortex using disposable Temno EvolutionR TT146 biopsy needle (Merit Medical Systems Inc., South Jordan, UT, USA). The biopsies were taken approximately 3 cm from the midline anterior to the coronal suture and were cylinder-shaped with a diameter of 5–7 mm and a length of 2–3 mm. Immunohistochemical analyses [22] on the biopsies have been described previously in detail. The process of selecting iNPH patients for shunting in Kuopio University Hospital has been described in previous studies [23]. Brain biopsies were analyzed with light microscopy by an experienced neuropathologist to determine cellular or neuronal immunoreactivity for amyloid-β (Aβ) and hyperphosphorylated tau (HPτ). The results were graded either as present or absent.

CSF biomarkers and apolipoprotein E

CSF specimens were collected from the patients before the shunt operation. CSF specimens included Aβ_1 - 42, total tau (T-tau), and phosphorylated tau (P-tau181). CSF specimens were available for 54 (87.1%) patients. Abnormal levels of CSF biomarkers have been studied extensively in correlation with cognitive impairment and neurodegenerative diseases [24]. The combination of CSF biomarker concentrations has been shown to help in separation iNPH from other cognitive disorders [25]. The patient’s APOE genotyping was previously conducted, as described in an earlier study [26], using a PCR method [27]. APOE genotype data was available for 59 (95.2%) patients. APOE has three isoforms in humans, APOE ɛ2, APOE ɛ3, and APOE ɛ4. They are expressed by their polymorphic alleles ɛ2, ɛ3, and ɛ4. Among the alleles ɛ4 is especially linked to diverse range of neurodegenerative diseases, with AD showing the strongest correlation [28]. Patients APOE genotypes were categorized into ɛ4 carriers or noncarriers.

Shunt response

Shunt responses for the iNPH registry patients are evaluated by measuring gait speed prior to shunt surgery, 3 months and 1 year after the shunt. An increase of 20% gait speed at 3-month follow-up visit is considered a positive shunt response. Data for pre-shunt and 3-month follow-up was available for 57 (91.9%) patients. Positive gait speed shunt response was reported with 30 (48.4%) of the patients at 3-month follow-up.

Ophthalmic examination

All the patients were examined by an experienced ophthalmologist (KK). Examinations were carried out within 4 months of the cognitive examination and were performed with a slit lamp biomicroscope and included a dilated fundus assessment. Both eyes were examined. Any clinical signs of pathology were pointed out. Applanation tonometry was used to measure IOP, except in two patients where rebound tonometry was used. Patients who had rebound tonometry taken had corneal thickness of 544/547μm (right/left eye) and 579/580μm. Higher corneal thickness may overestimate rebound tonometry measurements. The best corrected visual acuity (BCVA) was obtained during the visit using ETDRS charts. Spherical equivalent was calculated for both eyes, and it expresses the overall refractive error of an eye. Corneal thickness and axial length were measured by IOL master 700 (Carl Zeiss MeditecAG)

Visual field test

Study participants underwent standard automated perimetry (SAP) using the Swedish interactive threshold algorithm (SITA-Fast) 24-2 program with the Humphrey Field Analyzer II 750 (2010 Carl Zeiss Meditec). Visual field tests were performed by a trained ophthalmological nurse after correction of refractive errors. VF data was collected from the patients in the form of FP, FL, FN, MD, PSD, and VFI.

The patients were divided into two groups depending on their visual field test results. Patients who had FP rate ≥15% and/or FN≥25% in either eye were deemed unreliable (n = 19) while patients who had FP rate <15% and FN rate < 25% in both eyes were deemed as reliable (n = 43).

Ophthalmic imaging

SD-OCT was used to obtain data on macular thickness parameters, macular volume, and RNFL thickness, using the Spectralis S2610-C OCT (Software Version 6.9.4.0) device by Heidelberg Engineering, Germany. All the scans were performed by an experienced operator.

The macular thickness map was obtained by using the 3D macula protocol. The axial scans cover an area of 6×6 mm² in the macular region centered on the fovea as defined by the Early Treatment of Diabetic Retinopathy Study (ETDRS) [29]. The axial scans are composed of raster scans with a resolution of 768×496 (vertical×horizontal). Based on the ETDRS map, macula is divided into 3 concentric rings and 9 regions. Concentric rings measure 1 mm (innermost ring), 3 mm (inner ring) and 6 mm (outer ring) in diameter centered on the fovea. One region is in the innermost 1 mm ring centered on the fovea while the 3 mm ring and 6 mm ring are divided into four equal regions (Fig. 2). In the innermost ring, which includes the central thickness measurement defined as the average thickness of the 1 mm ring, a separate measurement was taken for the centermost point thickness. The OCT device measures the distance between the inner limiting membrane and the inner boundary of the retinal pigment epithelium to determine the macular thickness. Mean macular thickness gives the average macular thickness of the 9 regions in the ETDRS map.

Fig. 2

A) Macular thickness map in optical coherence tomography divided into 9 regions. OS, Outer Superior; IS, Inner Superior; C, Central; II, Inner Inferior; OI, Outer Inferior; OT, Outer Temporal; IT, Inner Temporal; IN, Inner Nasal; ON, Outer nasal. B) Concentric 1 mm, 3 mm, and 6 mm rings. C) Orientation of the optical coherence tomography. S, Superior; N, Nasal; I, Inferior; T, Temporal.

To measure RNFL thickness, we used SD-OCT for scan acquisition. Scanning protocol generates data over 6×6 mm² area centered on the optic disc. OCT device has a built-in automated algorithm, which identifies the center of the optic disc and places a 3.7 mm (12°) diameter circle around it. The system samples the circle to obtain an RNFL thickness map, which is divided into 7 sectors arranged in two concentric rings. The inner ring provides the global average (G) RNFL thickness, while the outer ring consists of six sectors that are equally divided. These six sectors are designated as temporal (T), temporal superior (TS), nasal superior (NS), nasal (N), nasal inferior (NI), and temporal inferior (TI).

Statistical analysis

The data were analyzed with Statistical Package for Social Sciences (IBM Corp. Released 2020. IBM SPSS Statistics for Windows, Version 27.0. Armonk, NY: IBM Corp). A p-value≤0.05 was considered statistically significant. To measure the significance of the differences between study participants with unreliable and reliable visual field test results against CERAD-NB and other variables we used independent T-test for continuous variables and Pearson Chi-square test for the non-continuous variables. For the independent T-test, Levene’s test for equality of variances was determined to evaluate difference between the variances. We calculated the effect size with Cohen’s d to determine the magnitude of differences between the tested groups and CERAD-NB. A Cohen’s d value of 0.8 or greater is generally considered to indicate a large effect, but effects should be assessed in the context of clinical significance, particularly when there is substantial overlap. Univariate Analysis of Variance with age as the covariate for each CERAD-NB subtest was performed.

RESULTS

Pairwise comparison of the demographics has been laid out in Tables 1–5. Unreliable and reliable groups were similar regarding gender and education. Age emerged as a significant variable between the two groups, revealing that the unreliable group were older than their counterparts (78.5 years versus 74.9 years in the reliable group, p = 0.049). There was no statistical significance between Aβ biopsy status and HPτ biopsy status, APOE allele status or CSF tau biomarkers in the two groups. The CSF Aβ₄₂ levels before the shunt differed between the groups, with the reliable group showing higher levels (745.5 pg/ml versus 584.8 pg/ml in the unreliable group, p = 0.012). The comparison of shunt response between the two groups, as evaluated by gait speed, revealed a similar profile with no significant differences observed (Table 1). Regarding the use of medication for comorbidities, the reliable group showed a higher percentage of patients using hypertension medications (76.7% versus 63.2% in the unreliable group). Conversely, the unreliable group had a higher percentage using diabetes medications (42.1% versus 27.9% in the reliable group). However, no statistically significant differences were found between the two groups.

Table 1

Comparison of multiple variables of interest in 62 iNPH patients based on their visual field test results, categorizing them as either reliable or unreliable

Variables	19 Unreliable			43 Reliable			Comparisons
	N	Mean±SD or %	Range	N	Mean±SD or %	Range	p	Cohen’s d
Sex (Male)		63.2			55.8		0.589
Education level (≤9 y)		57.9		42	61.9		0.767
Age (Eye examination)		78.5±6.3	68–90		74.9±6.4	60–87	0.049
Hypertension medication		63.2			76.7		0.269
Diabetes medication		42.1			27.9		0.270
Positive HPτ (AT8) in brain biopsy		10.5			2.3		0.165
Positive Aβ in brain biopsy		52.6			44.2		0.539
APOE positive allele ɛ4	18	16.7		41	29.3		0.306
CSF Aβ₄₂ before shunt (pg/ml)	15	584.8±214.5	98.0–918.6	39	745.5±199.3	364.0–1259.0	0.012
CSF Tau protein before shunt (pg/ml)	15	174.0±100.1	42.0–365.0	39	175.6±70.6	39.0–335.0	0.948
CSF Phosphorylated tau before shunt (pg/ml)	15	32.6±14.5	13.2–61.0	39	34.1±11.7	15.0–59.0	0.705
10MWT before shunt (s)		35.6±20.4	11.2–83.5	42	30.7±32.2	7.8–208.0	0.548
10MWT 3mo. post shunt (s)	17	24.1±13.5	6.5–46.5	41	24.7±24.3	6.4–149.0	0.923
10MWT 1y post shunt (s)	12	17.3±8.4	5.5–29.0	33	18.9±15.1	6.1–89.0	0.725
CERAD-NB Subtests
Verbal Fluency (animals in 60 s)	13.4±6.0	3–23		17.7±6.4	6–33	0.018	–0.672
15-Item Boston Naming Test (0–15 scale)	11.3±2.0	8–15		13.0±2.2	8–15	0.005	–0.795
Mini-Mental State Exam (MMSE, 0–30 scale)	22.4±4.0	14–27		25.6±3.2	15–30	0.002	–0.906
Constructional Praxis (0–11 scale)	7.4±1.4	4–10		9.1±1.7	7–11	<0.001	–1.066
Delayed Constructional Praxis (0–11 scale)	4.5±2.8	0–9		7.2±2.9	0–11	0.001	–0.939
Word List Learning (0–30 scale)	12.7±5.3	4–24		17.8±4.5	8–25	<0.001	–1.082
Word List Recall (0–10 scale)	3.4±2.4	0–8		5.4±2.2	1–9	0.002	–0.907
Word List Recall Savings %	57.1±38.3	0–120		80.2±26.4	20–175	0.024	–0.758
Word List Recognition % (0–100 scale)	84.7±11.5	55–100		91.6±8.5	65–100	0.011	–0.726
Clock Drawing (0–6 scale)	3.3±1.9	1–6		4.1±1.7	1–6	0.073	–0.502

[N] The number of participants if any data missing. Patients were categorized according to visual field indices as unreliable if they had false positive rate ≥15% and/or false negative rate ≥25% in either eye. Patients were categorized as reliable if they had false positive rate <15% and false negative rate <25% in both eyes. Chi square test was used for Sex, Education, Hypertension medication, Diabetes medication, HPτ, Aβ, and APOE, and independent T-tests for other variables. CSF, cerebrospinal fluid; HPτ, hyperphosphorylated tau; 10MWT, 10-meter walking test; CERAD-NB, Consortium to Establish a Registry for Alzheimer’s Disease – Neuropsychological Battery.

The unreliable group on average performed worse in all CERAD-NB subtests (Table 1). Nine out of the ten subtests proved to be statistically significant, Verbal Fluency, Boston Naming Test, MMSE, Constructional Praxis, Delayed Constructional Praxis, Word List Learning, Word List Recall, Word List Recall Savings and Word List Recognition. Only Clock Drawing did not show a statistically significant difference. There was a substantial amount of overlap in the range of subtests between the groups. For the nine subtests where significant differences were found, Cohen’s d effect sizes ranged from medium to large.

To assess age as a confounder for the CERAD-NB subtests between the groups, we performed a subsequent Univariate Analysis of Variance with age as a covariate. The impact of age as a covariate on the CERAD-NB subtests was found to be marginal (Table 2).

Table 2

Univariate Analysis of Variance with age as a covariate for CERAD-NB subtests between groups based on their visual field test results, categorizing them as either reliable or unreliable

CERAD -NB Subtests	Mean Difference	95% CI for difference	p
Verbal Fluency (animals in 60 s)	4.12	0.51–7.73	0.026
15-Item Boston Naming Test (0–15 scale)	1.38	0.19–2.57	0.024
Mini-Mental State Exam (MMSE, 0–30 scale)	3.03	1.05–5.02	0.003
Constructional Praxis (0–11 scale)	1.65	0.73–2.57	<0.001
Delayed Constructional Praxis (0–11 scale)	2.26	0.675–3.844	0.006
Word List Learning (0–30 scale)	4.90	2.18–7.63	<0.001
Word List Recall (0–10 scale)	1.87	0.59–3.16	0.005
Word List Recall Savings %	22.26	4.78–39.73	0.013
Word List Recognition % (0–100 scale)	6.02	0.64–11.40	0.029
Clock Drawing (0–6 scale)	0.72	–0.27–1.72	0.149

Patients were categorized according to visual field indices as unreliable if they had false positive rate ≥15% and/or false negative rate ≥25% in either eye. Patients were categorized as reliable if they had false positive rate <15% and false negative rate <25% in both eyes. CERAD-NB, Consortium to Establish a Registry for Alzheimer’s Disease – Neuropsychological Battery.

Most visual field parameters displayed statistically significant differences between the groups, except for FL in the left eye. Additionally, FPs in both the right and left eyes showed substantial differences between the groups. Similarly, FNs demonstrated significant disparities between the right and left eyes in the group comparison. The unreliable group received worse results in all the other tested visual field parameters compared to the reliable group, with significant statistical differences observed between the two groups in VFI, MD, and PSD in the right and left eyes (Table 3).

Table 3

Ophthalmological comparison of 62 iNPH patients based on their visual field test results, categorizing them as either reliable or unreliable

Variables	19 Unreliable patients			43 Reliable patients
	N	Mean±SD	Range	N	Mean±SD	Range	p
Intraocular pressure o.dx (mmHg)	18	14.3±2.8	8–18	35	14.9±3.1	10–20	0.485
Intraocular pressure o.sin (mmHg)	18	14.9±2.6	10–20	35	14.6±2.9	10–20	0.752
Corneal thickness o.dx (μm)	16	527.3±48.7	375–587	37	551.6±28.5	488–619	0.027
Corneal thickness o.sin (μm)	16	508.9±87.5	233–587	37	542.8±62.7	214–626	0.116
Eye axial length o.dx (mm)	17	23.49±1.33	22.13–27.96	39	23.62±1.67	21.67–30.67	0.775
Eye axial length o.sin (mm)	17	23.43±1.11	22.28–27.11	39	23.67±1.94	21.6–31.19	0.636
Visual acuity o.dx (ETDRS)		77.6±8.4	63–89		80.4±5.9	57–90	0.198
Visual acuity o.sin (ETDRS)		77.16±7.9	60–86		78.3±7.1	60–90	0.581
Spherical equivalent o.dx (D)		0.57±1.25	–1.25–3.13		0.26±1.68	–5.50–3.25	0.481
Spherical equivalent o.sin (D)		0.68±1.50	–1.38–2.88		0.090±1.59	–3.75–4.25	0.178
Visual field parameters
Visual field index o.dx (%)		66.8±31.3	8–99		91.1±9.3	60–100	0.003
Visual field index o.sin (%)		71.3±28.0	2–99	42	89.9±12.2	44–100	0.011
Mean deviation o.dx (Db)		–11.04±9.42	–28.77–0.00		–4.26±3.99	–18.70–0.72	0.007
Mean deviation o.sin (dB)		–10.77±9.11	–30.69–0.45	42	–4.88±4.65	–20.52–0.71	0.014
Pattern standard deviation o.dx (dB)		5.31±2.57	1.50–9.77		3.72±2.12	1.20–9.67	0.014
Pattern standard deviation o.sin (dB)		5.35±2.81	1.64–11.14	42	3.88±2.58	1.18–12.17	0.049
Fixation loss o.dx (%)		31.6±26.9	0–100		14.8±18.8	0–93	0.006
Fixation loss o.sin (%)		18.8±23.5	0–87	42	19.0±22.4	0–100	0.979
False positive o.dx (%)		7.3±7.6	0–23		2.6±2.9	0–12	0.016
False positive o.sin (%)		12.1±10.5	0–33	42	2.9±3.3	0–13	0.001
False negative o.dx (%)	16	19.8±20.6	3–75		8.2±6.4	0–23	0.041
False negative o.sin (%)	17	25.4±16.6	0–63	42	8.3±6.6	0–23	<0.001

[N] The number of participants if any data missing. Patients were categorized according to visual field indices as unreliable if they had false positive rate ≥15% and/or false negative rate ≥25% in either eye. Patients were categorized as reliable if they had false positive rate <15% and false negative rate <25% in both eyes. Independent T-test for equality of means. mmHg, millimeters of mercury; o.dx, right eye; o.sin, left eye; ETDRS, Early Treatment of Diabetic Retinopathy Study (chart); D, diopter; dB, decibel.

In the comparison of other ophthalmological variables between the unreliable and reliable groups, no clinically significant differences were found (Table 3). Corneal thickness on the right eye was flagged as statistically significant, but results are within range of normal references for Caucasian population [30] and bear no clinical significance. The axial length values for both groups were comparable and appear to be consistent with those of other populations with similar age and race demographics [31]. In our study, a significant proportion of patients in both the unreliable (57.9%) and reliable groups (37.2%) had undergone cataract surgery. Spherical equivalents were comparable between the groups without statistically significant difference.

When comparing macular thickness for the groups (Table 4), both eyes had a comparable profile, with only inner temporal thickness of the left eyes showing a statistical difference (p = 0.012). The reliable group’s left eyes had a mean inner temporal thickness of 333.8μm while the unreliable group had a mean thickness of 315.9μm. When comparing the right eyes, the unreliable group had a mean inner temporal thickness of 320.1μm, while the reliable group had a mean thickness of 318.7μm. Both values can be considered within normal range and have no clinical significance [32].

Table 4

The comparison of macular thickness [μm] between 62 iNPH patients based on their visual field test results, categorizing them as either reliable or unreliable

Thickness	Unreliable n = 19		Reliable n = 43
	Mean±SD	Range	Mean±SD	Range	p
Right eye
Outer superior	278.3±17.2	248–314	282.7±19.1	228–315	0.390
Inner superior	319.5±22.0	260–355	327.1±22.5	284–370	0.225
Central	286.6±26.6	245–367	280.1±33.0	198–346	0.455
Inner inferior	321.7±18.0	283–350	324.7±22.5	277–365	0.606
Outer inferior	273.6±21.5	235–323	276.7±18.6	245–320	0.568
Outer temporal	274.8±19.2	243–330	275.3±16.1	241–308	0.922
Inner temporal	320.1±18.5	282–367	318.7±21.6	269–368	0.807
Inner nasal	327.1±15.9	293–357	332.7±24.8	275–380	0.366
Outer nasal	293.2±17.6	260–322	297.5±19.8	263–335	0.415
Center	258.6±42.2	198–369	248.8±39.9	171–345	0.385
Left eye
Outer superior	279.1±15.0	259–318	286.0±17.8	248–321	0.143
Inner superior	315.4±17.0	270–343	327.0±26.0	243–372	0.081
Central	279.7±25.0	241–330	286.2±43.9	230–439	0.553
Inner inferior	312.7±21.6	268–341	325.3±23.9	268–391	0.054
Outer inferior	272.3±21.9	230–314	282.7±34.6	241–447	0.235
Outer temporal	291.5±17.0	260–329	299.0±22.3	247–365	0.200
Inner temporal	315.9±17.7	279–349	333.8±27.7	270–432	0.012
Inner nasal	309.9±21.9	267–354	322.8±24.2	279–385	0.052
Outer nasal	267.9±21.8	233–331	276.3±15.5	242–304	0.090
Center	260.4±39.4	196–334	251.8±47.1	188–435	0.492

Patients were categorized according to visual field indices as unreliable if they had false positive rate ≥15% and/or false negative rate ≥25% in either eye. Patients were categorized as reliable if they had false positive rate <15% and false negative rate <25% in both eyes. Central, central thickness of the 1 mm ring; Center, centermost point of the 1 mm ring. Independent T-test for equality of means. iNPH, Idiopathic Normal Pressure Hydrocephalus; SD, standard deviation.

The RNFL thickness map profiles were similar in both the unreliable and reliable groups (Table 5), and there were no statistically significant differences between them. Standard deviation of the data could be considered high. RNFL thickness in different populations has been studied extensively with multiple OCT machines. Normal RNFL thickness values vary greatly depending on factors like age, race, gender, axial length, spherical equivalent, and disc area. The RNFL thickness in a previous analysis of healthy individuals with a mean age of 53 was found to be 97.3μm, with a consistent thinning of approximately 1.5μm per decade of age [33]. The iNPH study population appears to have on average a thinner RNFL layer.

Table 5

The comparison of RNFL thickness [μm] between 62 iNPH patients based on their visual field test results, categorizing them as either reliable or unreliable

Thickness	19 Unreliable patients			43 Reliable patients
	N	Mean±SD	Range	N	Mean±SD	Range	p
Right eye
TS	18	116.1±21.6	59–143		123.0±33.0	16–168	0.415
NS	18	89.6±30.9	10–129		92.8±29.5	19–162	0.704
T	18	68.9±16.3	48–111		71.4±17.4	39–126	0.602
G	18	87.4±13.2	60–114		89.4±17.8	40–120	0.678
N	18	67.6±21.0	10–98		64.0±18.2	7–107	0.502
TI	18	116.3±23.4	72–161		128.9±28.1	54–187	0.099
NI	18	105.0±25.8	58–167		99.4±29.7	21–171	0.490
Left eye
TS	18	119.1±27.1	51–151	42	133.1±75.1	38–578	0.506
NS	18	99.4±30.6	48–184	42	109.5±60.9	23–455	0.446
T	17	64.8±11.9	42–89	42	74.5±30.1	29–226	0.845
G	17	85.6±12.5	58–108	42	94.3±26.0	32–218	0.197
N	18	70.2±33.6	27–192	42	68.8±20.5	22–159	0.203
TI	17	114.6±24.1	67–154	42	119.5±31.5	34–178	0.843
NI	18	107.7±34.5	64–196	42	105.9±31.6	3–158	0.566

[N] The number of participants if any data missing. Patients were categorized according to visual field indices as unreliable if they had false positive rate ≥15% and/or false negative rate ≥25% in either eye. Patients were categorized as reliable if they had false positive rate <15% and false negative rate <25% in both eyes. TS, temporal-superior; NS, nasal-superior; T, temporal; G, global average; N, nasal; TI, temporal-inferior; NI, nasal-inferior; SD, standard deviation. Independent T-test for equality of means.

DISCUSSION

Decline in cognitive capabilities examined via CERAD-NB subtests in iNPH patients associated well with visual field reliability indices. To our knowledge, there is a lack of studies investigating the relationship between CERAD-NB scores and visual field testing. It should be recommended that iNPH patients with decreased CERAD-NB scores who display potential ophthalmological findings be prioritized for examinations that rely less on cognitive abilities. Based on our study findings, we were unable to determine a CERAD-NB threshold that could be used to exclude iNPH patients from visual field testing due to significant overlap in subtest performance between patient groups. While the increase in age may contribute to the decline in cognitive abilities and the reliability of visual field tests, age alone should not serve as the sole reason to dismiss visual field examinations or their potential in screening patients for cognitive decline. Notably, older patients, including those in the reliable group, demonstrated the ability to provide reliable visual fields, with the oldest patient being 87 years old. Still, age plays a significant role in increasing the prevalence of iNPH and glaucoma, which is an important factor to consider when interpreting the results. Visual field testing is a resource-intensive and time-consuming examination. The annual costs of visual field tests are unknown, as there is a lack of data regarding the costs associated in Finland. The cost of a single visual field test at Kuopio University Hospital is 65€, with funding being allocated partly from the government and partly through outpatient clinic visit fees. Conducting multiple visual field tests on patients who are unable to provide reliable results constitutes a suboptimal use of limited healthcare resources. Therefore, there is a need to carefully evaluate the necessity and cost-effectiveness of visual field testing in specific patient populations, such as those with cognitive impairment.

Of the 62 patients evaluated for ophthalmic examinations, RNFL thickness and visual fields, one patient received a glaucoma diagnosis according to the Finnish Current Care Guidelines[34] during a scheduled follow-up visit six months later, prompted by suspicion of glaucoma during examinations. iNPH patients with NTG becomes exceedingly difficult to diagnose as patients might have normal intraocular pressure and unreliable visual fields. Even in the reliable group fourteen patients (32.6%) displayed VFI under 90% at least in one eye. Three of these patients showed RNFL thinning as well, but without correlation to visual field test results or clinical findings. Eleven of these patients had normal RNFL-thicknesses and physiological optic nerves. Reduced VFI findings appear to be at least partially attributed to cognitive impairment, which further supports the notion that visual field indices should be observed on a continuum rather than categorizing patients simply as reliable or unreliable based on certain threshold values. In our study population, increased rates of NTG were not found. RNFL thickness in iNPH patients may present with thinning, however it is important to exercise caution in drawing conclusions from these findings, as the lack of a comparable healthy control group measured with the same OCT device may hinder definitive interpretation. In our study population, RNFL thinning might be partly attributed to age and axial length. There is a strong correlation between axial length of the eye, myopia [35], and reduced RNFL thickness, especially with high myopic (Spherical equivalent≤–6 Diopter) individuals with axial length exceeding 25 mm [36]. The impact of standard deviation on the mean values, especially with a limited sample size, should be noted. Future research should aim to explore whether RNFL thinning in iNPH patients is consistently observed across multiple cohorts, particularly when compared to a suitable control group.

This study contains a few strong points, including a comprehensive and unique set of patient data involving multiple variables, which were obtained through an interdisciplinary collaboration between the ophthalmological and neurosurgical departments. It is important also to note some limitations in the study, including its cross-sectional design, lack of follow-up, and restriction to a single center. A statistical significance threshold set at p-value 0.05 can potentially lead to Type 1 errors with multiple comparisons. Due to the lack of a healthy comparison group, we compared our findings to those of previous studies where differences in methods should be noted. The timing between the acquisition of shunt and the ophthalmological and cognitive examinations varied within the study population. More recent NPH patients underwent both examinations and shunt procedures within months of each other, whereas earlier patients of the cohort from the NPH registry had received a shunt up to almost seven years earlier. Patients who received a shunt earlier might present with reduced susceptibility to other neurodegenerative diseases owing to prolonged shunt management. Conversely, they may be more susceptible to shunt-related complications compared to their counterparts. These factors have the potential to influence comparisons within the cohort and cause selection bias. Comorbidities were identified based on the use of medications for hypertension and diabetes, yet unexamined factors like environmental or lifestyle habits (e.g., smoking) could impact the study results.

In conclusion, iNPH patients with decreased CERAD-NB results produce less reliable visual field tests compared to their counterparts. Notably, these findings are independent of factors such as Aβ biopsy status, HPτ biopsy status, APOE allele status, CSF tau biomarkers, or the ophthalmological status of the patient. The reliable group displayed higher CSF Aβ₄₂ levels, which have been linked to normal cognition in individuals with brain amyloidosis [37]. Our patients demonstrated similar rates of Aβ biopsy status, supporting this finding. Our data does not support a threshold value for withholding a visual field test based on patients’ cognitive abilities. Importantly, our study raises a question: should patients who present unreliable visual field tests be screened for cognitive impairment?

AUTHOR CONTRIBUTIONS

Benjam Kemiläinen (Conceptualization; Data curation; Formal analysis; Investigation; Visualization; Writing – original draft; Writing – review & editing); Sonja Tiainen (Data curation; Investigation; Visualization; Writing – original draft); Tuomas Rauramaa (Resources); Antti J. Luikku (Resources; Visualization; Writing – review & editing); Sanna-Kaisa Herukka (Resources); Anne Koivisto (Resources); Mikko Hiltunen (Resources); Steven Verdoneer (Funding acquisition; Project administration); Ken Johnson (Project administration); Mieko Chambers (Project administration); Kai Kaarniranta (Conceptualization; Formal analysis; Funding acquisition; Investigation; Project administration; Resources; Supervision; Writing – review & editing); Ville Leinonen (Conceptualization; Formal analysis; Funding acquisition; Investigation; Project administration; Resources; Supervision; Writing – review & editing).

Footnotes

ACKNOWLEDGMENTS

We appreciate and acknowledge RN Marita Parviainen for performing the cognitive tests and upholding the iNPH registry and RN Tanja Jurvanen for performing ophthalmic imaging of the patients. ChatGPT 3.5, a language model developed by OpenAI in San Francisco, CA, USA, was utilized to assist in grammar correction.

FUNDING

The study was funded by Neurovision Imaging LLC, Kuopio University Hospital VTR Fund (5503770), Academy of Finland (333302, GeneCellNano Flagship), The Sigrid Juselius Foundation, The Päivikki and Sakari Sohlberg Foundation and The Finnish Eye Foundation.

CONFLICT OF INTEREST

Steven R. Verdooner is a founding member of Neurovision Imaging LLC. Mieko Chambers and Ken Johnson are employees of NeuroVision Imaging LLC.

All other authors have no conflict of interest to report.

DATA AVAILABILITY

The data supporting the results of this study can be obtained upon request from the corresponding author. These data are not accessible to the public due to privacy and ethical constraints.

References

Nakajima

, Yamada

, Miyajima

, Ishii

, Kuriyama

, Kazui

, Kanemoto

, Suehiro

, Yoshiyama

, Kameda

, Kajimoto

, Mase

, Murai

, Kita

, Kimura

, Samejima

, Tokuda

, Kaijima

, Akiba

, Kawamura

, Atsuchi

, Hirata

, Matsumae

, Sasaki

, Yamashita

, Aoki

, Irie

, Miyake

, Kato

, Mori

, Ishikawa

, Date

, Arai

, research committee of idiopathic normal pressure hydrocephalus (2021) Guidelines for Management of Idiopathic Normal Pressure Hydrocephalus (Third Edition): Endorsed by the Japanese Society of Normal Pressure Hydrocephalus. Neurol Med Chir (Tokyo) 61, 63–97.

Giordan

, Palandri

, Lanzino

, Murad

, Elder

(2018) Outcomes and complications of different surgical treatments for idiopathic normal pressure hydrocephalus: A systematic review and meta-analysis. J Neurosurg 131, 1024–1036.

Fasano

, Espay

, Tang-Wai

, Wikkelsö

, Krauss

(2020) Gaps, controversies, and proposed roadmap for research in normal pressure hydrocephalus. Mov Disord 35, 1945–1954.

Jaraj

, Rabiei

, Marlow

, Jensen

, Skoog

, Wikkelsø

(2014) Prevalence of idiopathic normal-pressure hydrocephalus. Neurology 82, 1449–1454.

Picascia

, Zangaglia

, Bernini

, Minafra

, Sinforiani

, Pacchetti

(2015) A review of cognitive impairment and differential diagnosis in idiopathic normal pressure hydrocephalus. Funct Neurol 30, 217–228.

Peterson

, Mole

, Keong

NCH

, DeVito

, Savulich

, Pickard

, Sahakian

(2019) Structural correlates of cognitive impairment in normal pressure hydrocephalus. Acta Neurol Scand 139, 305–312.

Nerg

, Junkkari

, Hallikainen

, Rauramaa

, Luikku

, Hiltunen

, Jääskeläinen

, Leinonen

, Hänninen

, Koivisto

(2021) The CERAD Neuropsychological Battery in patients with idiopathic normal pressure hydrocephalus compared with normal population and patients with mild Alzheimer’s disease. J Alzheimers Dis 81, 1117–1130.

Luikku

, Hall

, Nerg

, Koivisto

, Hiltunen

, Helisalmi

, Herukka

S-K

, Junkkari

, Sutela

, Kojoukhova

, Korhonen

, Mattila

, Lötjönen

, Rummukainen

, Alafuzoff

, Jääskeläinen

, Remes

, Solomon

, Kivipelto

, Soininen

, Rauramaa

, Leinonen

(2019) Predicting development of Alzheimer’s disease in patients with shunted idiopathic normal pressure hydrocephalus. J Alzheimers Dis 71, 1233–1243.

Fillenbaum

, van Belle

, Morris

, Mohs

, Mirra

, Davis

, Tariot

, Silverman

, Clark

, Welsh-Bohmer

, Heyman

(2008) Consortium to Establish a Registry for Alzheimer’s Disease (CERAD): The first twenty years. Alzheimers Dement 4, 96–109.

10.

Sotaniemi

, Pulliainen

, Hokkanen

, Pirttilä

, Hallikainen

, Soininen

, Hänninen

(2012) CERAD-neuropsychological battery in screening mild Alzheimer’s disease. Acta Neurol Scand 125, 16–23.

11.

Eleftheriou

, Huang-Link

, Lundin

(2021) Optical coherence tomography revealing ganglion cell loss in idiopathic normal pressure hydrocephalus. World Neurosurg 149, e1061–e1066.

12.

Killer

, Pircher

(2018) Normal tension glaucoma: Review of current understanding and mechanisms of the pathogenesis. Eye (Lond) 32, 924–930.

13.

Igarashi

, Honjo

, Asano

, Takagi

, Aihara

(2019) Optic disc cupping characteristics of normal pressure hydrocephalus patients with normal-tension glaucoma. Sci Rep 9, 3108.

14.

Hamarat

, Bartusis

, Deimantavicius

, Lucinskas

, Siaudvytyte

, Zakelis

, Harris

, Mathew

, Siesky

, Janulevicienė

, Ragauskas

(2021) Can the treatment of normal-pressure hydrocephalus induce normal-tension glaucoma? A narrative review of a current knowledge. Medicina (Kaunas) 57, 234.

15.

Phu

, Khuu

, Yapp

, Assaad

, Hennessy

, Kalloniatis

(2017) The value of visual field testing in the era of advanced imaging: Clinical and psychophysical perspectives. Clin Exp Optom 100, 313–332.

16.

Heijl

, Patella

, Bengtsson

(2021) Excellent Perimetry; the Field Analyzer Primer, 5th ed. Carl Zeiss Meditec, Inc.

17.

Yohannan

, Wang

, Brown

, Chauhan

, Boland

, Friedman

, Ramulu

(2017) Evidence-based criteria for assessment of visual field reliability. Ophthalmology 124, 1612–1620.

18.

Raman

, Khy Ching

, Sivagurunathan

, Ramli

, Mohd Khalid

(2019) The association between visual field reliability indices and cognitive impairment in glaucoma patients. J Glaucoma 28, 685–690.

19.

Cunha

, Pires

, Cruzeiro

, Almeida

ALM

, Martins

, Shigaeff

, Vale

(2022) Optical coherence tomography in neurodegenerative disorders. Arq Neuropsiquiatr 80, 180–191.

20.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43, 2412–2414.

21.

Relkin

, Marmarou

, Klinge

, Bergsneider

, Black

(2005) Diagnosing idiopathic normal-pressure hydrocephalus. Neurosurgery 57, S4–16; discussion ii-v.

22.

Seppälä

, Nerg

, Koivisto

, Rummukainen

, Puli

, Zetterberg

, Pyykkö

, Helisalmi

, Alafuzoff

, Hiltunen

, Jääskeläinen

, Rinne

, Soininen

, Leinonen

, Herukka

(2012) CSF biomarkers for Alzheimer disease correlate with cortical brain biopsy findings. Neurology 78, 1568–1575.

23.

Junkkari

, Luikku

, Danner

, Jyrkkänen

, Rauramaa

, Korhonen

, Koivisto

, Nerg

, Kojoukhova

, Huttunen

, Jääskeläinen

, Leinonen

(2019) The Kuopio idiopathic normal pressure hydrocephalus protocol: Initial outcome of 175 patients. Fluids Barriers CNS 16, 21.

24.

Hampel

, Blennow

(2004) CSF tau and β-amyloid as biomarkers for mild cognitive impairment. Dialogues Clin Neurosci 6, 379–390.

25.

Jeppsson

, Wikkelsö

, Blennow

, Zetterberg

, Constantinescu

, Remes

, Herukka

S-K

, Rauramaa

, Nagga

, Leinonen

, Tullberg

(2019) CSF biomarkers distinguish idiopathic normal pressure hydrocephalus from its mimics. J Neurol Neurosurg Psychiatry 90, 1117–1123.

26.

Pyykkö

, Helisalmi

, Koivisto

, Mölsä

JAA

, Rummukainen

, Nerg

, Alafuzoff

, Savolainen

, Soininen

, Jääskeläinen

, Rinne

, Leinonen

, Hiltunen

(2012) APOE4 predicts amyloid-β in cortical brain biopsy but not idiopathic normal pressure hydrocephalus. J Neurol Neurosurg Psychiatry 83, 1119–1124.

27.

Tsukamoto

, Watanabe

, Matsushima

, Kinoshita

, Kato

, Hashimoto

, Kurokawa

, Teramoto

(1993) Determination by PCR-RFLP of apo E genotype in a Japanese population. J Lab Clin Med 121, 598–602.

28.

Yin

, Wang

(2018) ApoE and neurodegenerative diseases in aging. In Aging and Aging-Related Diseases: Mechanisms and Interventions, Wang

, ed. Springer, Singapore, pp. 77–92.

29.

Hee

, Puliafito

, Duker

, Reichel

, Coker

, Wilkins

, Schuman

, Swanson

, Fujimoto

(1998) Topography of diabetic macular edema with optical coherence tomography. Ophthalmology 105, 360–370.

30.

Galgauskas

, Juodkaite

, Tutkuvienė

(2014) Age-related changes in central corneal thickness in normal eyes among the adult Lithuanian population. Clin Interv Aging 9, 1145–1151.

31.

Lee

, Klein

BEK

, Klein

, Quandt

, Wong

(2009) Age, stature, and education associations with ocular dimensions in an older white population. Arch Ophthalmol 127, 88–93.

32.

Grover

, Murthy

, Brar

, Chalam

(2009) Normative data for macular thickness by high-definition spectral-domain optical coherence tomography (spectralis). Am J Ophthalmol 148, 266–271.

33.

Alasil

, Wang

, Keane

, Lee

, Baniasadi

, de Boer

, Chen

(2013) Analysis of normal retinal nerve fiber layer thickness by age, sex, and race using spectral domain optical coherence tomography. J Glaucoma 22, 532–541.

34.

Leinonen

, Harju

, Hagman

, Honkamo

, Marttila

, Määttä

, Saarela

, Vaajanen

, Vesti

, Komulainen

(2024) The Finnish current care guideline for open-angle glaucoma. Acta Ophthalmol 102, 151–171.

35.

Meng

, Butterworth

, Malecaze

, Calvas

(2011) Axial length of myopia: A review of current research. Ophthalmologica 225, 127–134.

36.

Rauscher

, Sekhon

, Feuer

, Budenz

(2009) Myopia affects retinal nerve fiber layer measurements as determined by optical coherence tomography. J Glaucoma 18, 501–505.

37.

Sturchio

, Dwivedi

, Young

, Malm

, Marsili

, Sharma

, Mahajan

, Hill

, Andaloussi

, Poston

, Manfredsson

, Schneider

, Ezzat

, Espay

(2021) High cerebrospinal amyloid-β 42 is associated with normal cognition in individuals with brain amyloidosis. EClinicalMedicine 38, 100988.