Cerebrospinal Fluid Biomarkers in the Differential Diagnosis of Normal Pressure Hydrocephalus

Oja K¹, Lombardi V², Benesova T³

Abstract

Background: Idiopathic normal pressure hydrocephalus (iNPH) is a treatable cause of gait disturbance, cognitive impairment and urinary incontinence in older adults, but overlaps clinically with Alzheimer's disease (AD). Objective: To review evidence on cerebrospinal fluid (CSF) biomarkers used to differentiate iNPH from AD and predict shunt-responsiveness. Methods: Narrative review of diagnostic-criteria papers, biopsy-correlation studies, biomarker case-control series, meta-analyses and shunt-outcome cohorts. Findings: iNPH shows low amyloid-β42 with normal or modestly elevated tau/phospho-tau, contrasting with the low amyloid-β42/high tau pattern of AD; phospho-tau distinguishes the two with high sensitivity and specificity. Leucine-rich alpha-2-glycoprotein, neurofilament light chain and L1CAM add value. CSF tap testing retains higher specificity than any single biomarker for predicting shunt response. Conclusions: No single CSF biomarker substitutes for dynamic CSF testing, but biomarker panels usefully flag comorbid neurodegeneration and refine patient counselling.

Keywords: normal pressure hydrocephalus; cerebrospinal fluid biomarkers; amyloid-beta; tau protein; shunt responsiveness

Introduction

Idiopathic normal pressure hydrocephalus (iNPH) is a communicating hydrocephalus of older adults presenting with the classic Hakim triad of gait disturbance, cognitive impairment and urinary incontinence, occurring with normal opening pressure and ventriculomegaly out of proportion to cortical atrophy [1]. It is among the few potentially reversible causes of dementia-like presentations, with cerebrospinal fluid (CSF) shunting producing meaningful improvement in a substantial proportion of well-selected patients, as confirmed most recently by a placebo-controlled randomized trial [2]. International and Japanese guidelines have refined diagnostic criteria and incorporated imaging findings such as disproportionately enlarged subarachnoid-space hydrocephalus (DESH) into structured algorithms [3,4].

Despite this progress, patient selection remains difficult. The clinical triad is nonspecific in the elderly, and Alzheimer's disease (AD) is the most important differential and the most frequent comorbidity: cortical biopsy series report AD-consistent amyloid and tau pathology in roughly one-fifth to two-thirds of iNPH patients undergoing shunt surgery [5,6,7]. Distinguishing "pure" iNPH from iNPH with comorbid AD, and from AD alone misread as hydrocephalus because of age-related ventriculomegaly, matters clinically because it shapes expectations about which symptom domains (typically gait more than cognition) are likely to improve after shunting, and because unnecessary surgery in patients whose decline is driven predominantly by neurodegeneration exposes them to procedural risk without commensurate benefit. CSF biomarkers, already validated for AD diagnosis, have therefore been investigated extensively as adjuncts for differentiating iNPH from AD and for predicting which patients will benefit from shunting. This review synthesises that literature, covering diagnostic guidelines, the core amyloid/tau panel, emerging candidate biomarkers, the relationship between biomarkers and CSF dynamic testing, and dedicated shunt-outcome studies.

Diagnostic Challenge of Overlapping Neurodegeneration

Gait disturbance, cognitive slowing and urinary symptoms are common in both iNPH and AD, and neuroimaging alone cannot reliably separate the two, particularly when ventriculomegaly and cortical atrophy coexist. Early series established that concomitant AD pathology, identified at cortical biopsy during shunt placement, is present in a meaningful minority of clinically diagnosed iNPH patients, though its presence does not uniformly predict a worse shunt outcome when iNPH is accurately diagnosed [5]. A larger biopsy cohort of 468 patients confirmed that amyloid and tau histopathology carry independent long-term prognostic value beyond the clinical syndrome [7], and a more recent series found AD-consistent pathology in roughly 19% of 142 shunted patients, correlating biopsy findings with high-volume lumbar puncture results [6]. This overlap motivates the search for a fluid biomarker that can flag comorbid AD without requiring invasive biopsy, since misattributing AD-driven decline to reversible hydrocephalus—or vice versa—has direct consequences for surgical candidacy and prognostic counselling.

CSF Amyloid and Tau Biomarkers

The core AD CSF signature—reduced amyloid-β42 (Aβ42) with elevated total tau (t-tau) and phosphorylated tau (p-tau)—is well established. Early work in iNPH found a partially overlapping but distinguishable pattern: Aβ42 is reduced in iNPH nearly as much as in AD, but t-tau and p-tau are typically normal or only mildly elevated, contrasting with the marked tau elevation of AD [8,9]. A systematic review and meta-analysis pooling ten studies (413 iNPH, 186 AD, 147 controls) confirmed this pattern: relative to AD, p-tau achieved the best discrimination (sensitivity 94.3%, specificity 85.1%), followed by t-tau (82.8%/84.2%), while Aβ42 alone performed poorly (81.3%/50.6%) because it is low in both conditions [10]. Subsequent systematic reviews of iNPH versus AD and subcortical ischemic vascular disease reached concordant conclusions, with t-tau and p-tau—rather than Aβ42—carrying most of the discriminatory weight [11,12].

An important caveat is that hydrocephalus itself appears to suppress CSF concentrations of amyloid precursor protein fragments and tau, plausibly through compression of the interstitial space and impaired protein clearance; these levels rise after shunting decompresses the brain, meaning a single preoperative sample may understate true neurodegeneration and complicate cross-sectional comparison with AD [13]. This "global suppression" hypothesis explains why an isolated low Aβ42—the single most consistent iNPH finding—is a poor discriminator of AD in this population despite being highly informative in patients without hydrocephalus, and argues for interpreting biomarkers as ratios (Aβ42/Aβ40, p-tau/Aβ42) rather than as absolute concentrations. Consistent with this pathophysiological account, a large biomarker cohort (the Bologna Pro-Hydro study) found iNPH patients had significantly lower Aβ42, Aβ40, t-tau and p-tau than controls, alongside elevated neurofilament light chain—an axonal-injury marker not suppressed by the same mechanism [14]. A comparison of iNPH against multiple neurodegenerative dementias similarly found higher Aβ42 and lower t-tau/p-tau in iNPH than in AD and Lewy body dementia [15]. Overall, the Aβ42/tau pattern is useful chiefly as a ratio or combination rather than through any single analyte in isolation, and clinicians should be cautious about applying standard AD cutoffs, derived in non-hydrocephalic populations, directly to iNPH CSF.

Other Candidate CSF Biomarkers

Beyond the core AD panel, several iNPH-enriched biomarkers have been proposed. Leucine-rich alpha-2-glycoprotein (LRG), discovered through CSF proteomics, is specifically elevated in iNPH; combining an LRG cutoff of 67 ng/mL with a tau cutoff of 200 pg/mL achieved 81.6% sensitivity and 78.6% specificity for shunt responsiveness in one cohort, and LRG levels were significantly higher in shunt responders than non-responders (96.8 versus 29.2 ng/mL) [16]. A follow-up study confirmed elevated CSF LRG alongside the expected reduction in AD-type biomarkers, supporting LRG's utility independent of comorbid amyloid pathology [17]. Mass-spectrometry-based proteomic panels have additionally implicated L1 cell adhesion molecule (L1CAM) and other proteins that distinguish iNPH from both controls and AD [18], and a large-scale unbiased CSF proteomic analysis has begun to define molecular signatures specific to iNPH, offering candidate biomarkers beyond the classical panel [19].

Neurofilament light chain (NfL), a marker of axonal injury, is consistently elevated in iNPH relative to controls, though not specifically versus other neurodegenerative conditions; in a large series distinguishing iNPH from 297 non-iNPH "mimics," a combination of t-tau, Aβ40 and monocyte chemoattractant protein-1 (MCP-1)—reflecting astrocyte activation—separated iNPH from cognitive and movement-disorder mimics with good accuracy [20]. NfL also carries prognostic weight: higher preoperative NfL and t-tau were associated with less favourable outcome after shunt surgery in a cohort of 455 patients, although most patients with elevated markers still improved [21]. Collectively, these biomarkers suggest that no single analyte captures iNPH pathophysiology, but multi-marker panels combining amyloid, tau, and markers of axonal or astrocytic injury outperform any individual biomarker.

CSF Dynamic Testing (Tap Test/Lumbar Drainage) versus Biomarkers

Dynamic CSF testing remains the clinical mainstay for predicting shunt response. The high-volume lumbar tap test (removal of 30–50 mL CSF with pre/post gait and cognitive assessment) and prolonged external lumbar drainage (ELD) were validated in a prospective study of 151 patients, in which 66% improved after ELD and 84% of ELD-positive patients subsequently improved after shunting [22]. A dedicated analysis of tap-test performance across grading scales found sensitivity and specificity that varied considerably by outcome measure and threshold [23], and a formal cost-effectiveness and accuracy analysis concluded that ELD approaches but does not reach ideal sensitivity and specificity, and that no single screening test—biomarker or dynamic—is definitive on its own [24].

Compared against dynamic testing, individual CSF biomarkers generally show lower specificity for shunt response, but they carry two advantages: they are obtainable from the same single lumbar puncture used diagnostically, avoiding a separate prolonged drainage admission with its associated bed-days and small risk of meningitis, and they can flag comorbid AD pathology that dynamic testing cannot detect. The complementary, rather than competing, relationship between the two approaches is illustrated by the LRG-plus-tau-test combination, in which biomarker status stratified tap-test-positive patients into four groups with markedly different rates of postoperative cognitive improvement, from 100% of shunt responders in the low-tau/high-LRG group to only 33% in the high-tau/low-LRG group [16]. This kind of stratification suggests biomarkers are best positioned as a triage or risk-refinement layer applied on top of, rather than instead of, dynamic testing.

Predicting Shunt Responsiveness

Beyond simple case-control differentiation, several studies have directly examined which CSF biomarkers predict postoperative outcome. A study of ventricular CSF found that a higher phosphorylated-tau/Aβ42 ratio correlated with cortical AD pathology at biopsy and with less favourable long-term shunt outcome, though the authors cautioned against using this ratio alone to exclude patients from surgery [25]. A systematic review and meta-analysis of 13 studies spanning 37 unique CSF and venous biomarkers found that lumbar p-tau and t-tau were significantly higher in shunt non-responders, but that no single biomarker achieved both high sensitivity and specificity for predicting response, underscoring the continued primacy of dynamic testing [26]. A prospective cohort measuring Aβ1-40, Aβ1-42, t-tau, p-tau and NfL found that lower levels of each biomarker individually predicted better long-term gait outcome after shunting, reinforcing the concept that a "cleaner" neurodegeneration profile predicts better surgical response [27]. The 2025 PENS trial, the first placebo-controlled randomized trial of shunting in iNPH, confirmed that shunt surgery significantly improves gait and reduces falls specifically among patients selected via positive response to CSF drainage testing, reinforcing dynamic testing's centrality even as biomarker research matures [2].

Table 2: Selected studies relating CSF biomarkers or dynamic testing to shunt-responsiveness prediction in iNPH [22,16,25,26,27,2].

Study (year) Design / n Predictor(s) Key finding

evaluated

Marmarou Prospective, External lumbar 66% improved after 2005 n=151 drainage (ELD) ELD; 84% of ELD-positive patients improved after shunting Nakajima n=90 (tap Tap test + CSF Combined strategy: 2011 test); 52 LRG/tau 81.6% sensitivity, shunted 78.6% specificity for shunt response Patel 2012 Ventricular p-tau/Aβ42 ratio Higher ratio CSF cohort correlated with AD pathology at biopsy and less favourable outcome Thavarajasi Systematic CSF and venous Lumbar p-tau and ngam 2022 review, 13 biomarkers t-tau higher in studies, 37 non-responders; no biomarkers single biomarker sufficiently accurate alone Study (year) Design / n Predictor(s) Key finding

evaluated

Darrow 2022 Prospective Aβ1-40, Aβ1-42, Lower levels of each cohort t-tau, p-tau, NfL biomarker individually predicted better long-term gait outcome Luciano Randomized Response to CSF Shunting significantly 2025 (PENS , placebo-co drainage testing improved gait and trial) ntrolled (selection reduced falls in criterion) drainage-responsive patients

Current Limitations and Research Directions

Substantial heterogeneity across studies—in CSF sampling site (lumbar versus ventricular), assay platform, patient selection, and outcome definitions—limits direct comparison and explains some contradictory findings, particularly for t-tau and p-tau associations with outcome [26]. The biological confound that hydrocephalus itself suppresses baseline biomarker levels, normalising partially after shunting, complicates the interpretation of any single preoperative measurement as a pure marker of neurodegeneration [13]. Emerging unbiased proteomic approaches may identify iNPH-specific signatures less confounded by coexisting AD pathology [19], while refined quantitative imaging biomarkers such as automated DESH scoring offer a complementary, non-invasive axis for multimodal prediction models that combine imaging, dynamic testing and CSF biomarkers [28]. Prospective, adequately powered studies using standardized biomarker panels and harmonized outcome measures, ideally nested within trials such as the PENS study, are needed before any CSF biomarker can be recommended as a standalone diagnostic or prognostic test.

Conclusion

CSF biomarkers meaningfully inform, but do not replace, the differential diagnosis and shunt-responsiveness prediction of iNPH. The combination of reduced Aβ42 with relatively preserved tau distinguishes iNPH from AD with reasonable accuracy, especially using p-tau, while iNPH-enriched markers such as LRG and injury markers such as NfL add prognostic value. Dynamic CSF testing retains higher specificity for shunt response than any single biomarker, and current evidence supports using CSF biomarkers as an adjunct—for flagging comorbid neurodegeneration and refining patient counselling—rather than as a substitute for tap testing or lumbar drainage in surgical decision-making.

References

1. Relkin, N., Marmarou, A., Klinge, P., Bergsneider, M., Black, P.M. (2005) Diagnosing idiopathic normal-pressure hydrocephalus. Neurosurgery, 57(3 Suppl): S4-S16.
2. Luciano, M.G., Williams, M.A., Hamilton, M.G., et al. (2025) A Randomized Trial of Shunting for Idiopathic Normal-Pressure Hydrocephalus. New England Journal of Medicine, 393(22): 2198-2209.
3. Nakajima, M., Yamada, S., Miyajima, M., Ishii, K., Kuriyama, N., Kazui, H., Kanemoto, H., Suehiro, T., Yoshiyama, K., Kameda, M., et al. (2021) Guidelines for Management of Idiopathic Normal Pressure Hydrocephalus (Third Edition): Endorsed by the Japanese Society of Normal Pressure Hydrocephalus. Neurologia Medico-chirurgica, 61(2): 63-97.
4. Hamilton, M.G., Williams, M.A., Edwards, S., Tullberg, M. (2025) Guidelines for Diagnosis and Management of Idiopathic Normal Pressure Hydrocephalus. Neurosurgery Clinics of North America, 36(2): 199-205.
5. Golomb, J., Wisoff, J., Miller, D.C., Boksay, I., Kluger, A., Weiner, H., Salton, J., Graves, W. (2000) Alzheimer's disease comorbidity in normal pressure hydrocephalus: prevalence and shunt response. Journal of Neurology, Neurosurgery and Psychiatry, 68(6): 778-781.
6. Pomeraniec, I.J., Bond, A.E., Lopes, M.B., Jane, J.A. Sr. (2016) Concurrent Alzheimer's pathology in patients with clinical normal pressure hydrocephalus: correlation of high-volume lumbar puncture results, cortical brain biopsies, and outcomes. Journal of Neurosurgery, 124(2): 382-388.
7. Leinonen, V., Koivisto, A.M., Alafuzoff, I., Pyykkö, O.T., Rummukainen, J., von Und Zu Fraunberg, M., Jääskeläinen, J.E., Soininen, H., Rinne, J., Savolainen, S. (2012) Cortical Brain Biopsy in Long-Term Prognostication of 468 Patients with Possible Normal Pressure Hydrocephalus. Neurodegenerative Diseases, 10(1-4): 166-169.
8. Kapaki, E.N., Paraskevas, G.P., Tzerakis, N.G., Sfagos, C., Seretis, A., Kararizou, E., Vassilopoulos, D. (2007) Cerebrospinal fluid tau, phospho-tau181 and beta-amyloid1-42 in idiopathic normal pressure hydrocephalus: a discrimination from Alzheimer's disease. European Journal of Neurology, 14(2): 168-173.
9. Jeppsson, A., Zetterberg, H., Blennow, K., Wikkelsö, C. (2013) Idiopathic normal-pressure hydrocephalus: pathophysiology and diagnosis by CSF biomarkers. Neurology, 80(15): 1385-1392.
10. Chen, Z., Liu, C., Zhang, J., Relkin, N., Xing, Y., Li, Y. (2017) Cerebrospinal fluid Aβ42, t-tau, and p-tau levels in the differential diagnosis of idiopathic normal-pressure hydrocephalus: a systematic review and meta-analysis. Fluids and Barriers of the CNS, 14(1): 13.
11. Manniche, C., Hejl, A.M., Hasselbalch, S.G., Simonsen, A.H. (2019) Cerebrospinal Fluid Biomarkers in Idiopathic Normal Pressure Hydrocephalus versus Alzheimer's Disease and Subcortical Ischemic Vascular Disease: A Systematic Review. Journal of Alzheimer's Disease, 68(1): 267-279.
12. Manniche, C., Simonsen, A.H., Hasselbalch, S.G., Andreasson, U., Zetterberg, H., Blennow, K., Høgh, P., Juhler, M., Hejl, A.M. (2020) Cerebrospinal Fluid Biomarkers to Differentiate Idiopathic Normal Pressure Hydrocephalus from Subcortical Ischemic Vascular Disease. Journal of Alzheimer's Disease, 75(3): 937-947.
13. Graff-Radford, N.R. (2014) Alzheimer CSF biomarkers may be misleading in normal-pressure hydrocephalus. Neurology, 83(17): 1573-1575.
14. Abu-Rumeileh, S., Giannini, G., Polischi, B., Albini-Riccioli, L., Milletti, D., Oppi, F., Stanzani-Maserati, M., Capellari, S., Mantovani, P., Palandri, G., Cortelli, P., Cevoli, S., Parchi, P. (2019) Revisiting the Cerebrospinal Fluid Biomarker Profile in Idiopathic Normal Pressure Hydrocephalus: The Bologna Pro-Hydro Study. Journal of Alzheimer's Disease, 68(2): 723-733.
15. Said, H.M., Kaya, D., Yavuz, I., Dost, F.S., Altun, Z.S., Isik, A.T. (2022) A Comparison of Cerebrospinal Fluid Beta-Amyloid and Tau in Idiopathic Normal Pressure Hydrocephalus and Neurodegenerative Dementias. Clinical Interventions in Aging, 17: 467-477.
16. Nakajima, M., Miyajima, M., Ogino, I., Watanabe, M., Miyata, H., Karagiozov, K.L., Arai, H., Hagiwara, Y., Segawa, T., Kobayashi, K., Hashimoto, Y. (2011) Leucine-rich α-2-glycoprotein is a marker for idiopathic normal pressure hydrocephalus. Acta Neurochirurgica, 153(6): 1339-1346.
17. Vanninen, A., Nakajima, M., Miyajima, M., Rauramaa, T., Kokki, M., Musialowicz, T., Mäkinen, P.M., Herukka, S.K., Koivisto, A.M., Jääskeläinen, J.E., Hiltunen, M., Leinonen, V. (2021) Elevated CSF LRG and Decreased Alzheimer's Disease Biomarkers in Idiopathic Normal Pressure Hydrocephalus. Journal of Clinical Medicine, 10(5): 1105.
18. Torretta, E., Arosio, B., Barbacini, P., Capitanio, D., Rossi, P.D., Moriggi, M., Clerici, M., Mari, D., Cesari, M., Gelfi, C. (2021) Novel Insight in Idiopathic Normal Pressure Hydrocephalus (iNPH) Biomarker Discovery in CSF. International Journal of Molecular Sciences, 22(15): 8034.
19. Kamalian, A., Shirzadeh Barough, S., Ho, S.G., et al. (2024) Molecular signatures of normal pressure hydrocephalus: a large-scale proteomic analysis of cerebrospinal fluid. Fluids and Barriers of the CNS, 21: 64.
20. Jeppsson, A., Wikkelsö, C., Blennow, K., Zetterberg, H., Constantinescu, R., Remes, A.M., Herukka, S.K., Rauramaa, T., Nagga, K., Leinonen, V., Tullberg, M. (2019) CSF biomarkers distinguish idiopathic normal pressure hydrocephalus from its mimics. Journal of Neurology, Neurosurgery and Psychiatry, 90(10): 1117-1123.
21. Braun, M., Bjurnemark, C., Seo, W., et al. (2022) Higher levels of neurofilament light chain and total tau in CSF are associated with negative outcome after shunt surgery in patients with normal pressure hydrocephalus. Fluids and Barriers of the CNS, 19: 15.
22. Marmarou, A., Young, H.F., Aygok, G.A., et al. (2005) Diagnosis and management of idiopathic normal-pressure hydrocephalus: a prospective study in 151 patients. Journal of Neurosurgery, 102(6): 987-997.
23. Ishikawa, M., Hashimoto, M., Mori, E., Kuwana, N., Kazui, H. (2012) The value of the cerebrospinal fluid tap test for predicting shunt effectiveness in idiopathic normal pressure hydrocephalus. Fluids and Barriers of the CNS, 9(1): 1.
24. Burnett, M.G., Sonnad, S.S., Stein, S.C. (2006) Screening tests for normal-pressure hydrocephalus: sensitivity, specificity, and cost. Journal of Neurosurgery, 105(6): 823-829.
25. Patel, S., Lee, E.B., Xie, S.X., Law, A., Jackson, E.M., Arnold, S.E., Clark, C.M., Shaw, L.M., Grady, M.S., Trojanowski, J.Q., Hamilton, R.H. (2012) Phosphorylated tau/amyloid beta 1-42 ratio in ventricular cerebrospinal fluid reflects outcome in idiopathic normal pressure hydrocephalus. Fluids and Barriers of the CNS, 9: 7.
26. Thavarajasingam, S.G., El-Khatib, M., Vemulapalli, K.V., Iradukunda, H.A.S., Laleye, J., Russo, S., Eichhorn, C., Eide, P.K. (2022) Cerebrospinal fluid and venous biomarkers of shunt-responsive idiopathic normal pressure hydrocephalus: a systematic review and meta-analysis. Acta Neurochirurgica, 164(7): 1719-1746.
27. Darrow, J.A., Lewis, A., Gulyani, S., Khingelova, K., Rao, A., Wang, J., Zhang, Y., Luciano, M., Yasar, S., Moghekar, A. (2022) CSF Biomarkers Predict Gait Outcomes in Idiopathic Normal Pressure Hydrocephalus. Neurology: Clinical Practice, 12(2): 91-101.
28. Behndig, S., Lalou, A., Axelsson, J., et al. (2025) qDESH: a method to quantify disproportionately enlarged subarachnoid space hydrocephalus. Fluids and Barriers of the CNS, 22: 67.