Cerebrospinal fluid Aβ42, t-tau, and p-tau levels in the differential diagnosis of idiopathic normal-pressure hydrocephalus: a systematic review and meta-analysis

Abstract

Objectives: The purpose of this systematic review and meta-analysis was to evaluate the performance of cerebrospinal fluid (CSF) beta amyloid 42 (Aβ42), total tau (t-tau), and phosphorylated tau (p-tau) as potential diagnostic biomarkers for idiopathic normal-pressure hydrocephalus (iNPH) and to assess their utility indistinguishing patients with iNPH from those with Alzheimer disease (AD) and healthy normal controls.

Methods: Studies were identified by searching PubMed, Embase, the Cochrane Library, Web of Science, Chinese National Knowledge Infrastructure (CNKI), Wanfang Chinese Periodical Database, VIP Chinese database, and Chinese Bio-medicine Database (CBM) before August 2016. The standardized mean difference (SMD) and 95% confidence interval (CI), comparing CSF Aβ42, t-tau, and p-tau levels between iNPH, AD and healthy controls, were calculated using random-effects models. Subgroup analyses were created according to ethnicity (Caucasian or Asian) and CSF type (lumbar or ventricular), and the publication bias was estimated using Egger's test and the Begg's test.

Results: A total of 10 studies including 413 patients with iNPH, 186 patients with AD and 147 healthy controls were included in this systematic review and meta-analysis. The concentrations of CSF t-tau, and p-tau were significantly lower in iNPH patients compared to AD (SMD = -1.26, 95% CI -1.95 to -0.57, P = 0.0004; SMD = -1.54, 95% CI -2.34 to -0.74, P = 0.0002, respectively) and lower than healthy controls (SMD = -0.80, 95% CI -1.50 to -0.09, P = 0.03; SMD = -1.12, 95% CI -1.38 to -0.86, P < 0.00001, respectively). Patients with iNPH had significantly lower Aβ42 levels compared with controls (SMD = -1.14, 95% CI -1.74 to -0.55, P = 0.0002), and slightly higher Aβ42 levels compared with AD patients (SMD = 0.32, 95% CI 0.00-0.63, P = 0.05). Subgroup analyses showed that the outcomes may have been influenced by ethnicity and CSF source. Compared to AD, overall sensitivity in differentiating iNPH was 0.813 (95% CI 0.636-0.928) for Aβ42, 0.828 (95% CI 0.732-0.900) for t-tau, 0.943 (95% CI 0.871-0.981) for p-tau. Relative to AD, overall specificity in differentiating iNPH was 0.506 (95% CI 0.393-0.619) for Aβ42, 0.842 (95% CI 0.756-0.907) for t-tau, 0.851 (95% CI 0.767-0.914) for p-tau.

Conclusion: The results of our meta-analysis suggest that iNPH may be associated with significantly reduced levels of CSF Aβ42, t-tau and p-tau compared to the healthy normal state. Compared to AD, both t-tau and p-tau were significantly decreased in iNPH, but CSF Aβ42 was slightly increased. Prospective studies are needed to further assess the clinical utility of these and other CSF biomarkers in assisting in the diagnosis of iNPH and differentiating it from AD and other neurodegenerative disorders.

Keywords: Alzheimer’s disease; CSF biomarkers; Idiopathic normal-pressure hydrocephalus; Meta-analysis; Systematic review.

Fig. 1

Flow chart of the search and selection process

Fig. 2

Sensitivity analysis of CSF Aβ42 levels in iNPH compared to AD. Sensitivity analysis was performed by removing studies one by one and comparing the pooled estimate from the remaining studies with the pooled estimate from all studies. The X axis represents the pooled estimate and the Y axis represents the excluded studies. Analyses were conducted on the raw data and the pooled estimate was 0.33 (95% CI 0.00–0.65). The pooled estimate was different when the study of Seppala et al. [26], Miyajima et al. [24], Lim et al. [27], Pyykko et al. [21], Schirinzi et al. [28] were omitted separately. The pooled estimates (95% CI) were 0.37 (95% CI −0.05 to 0.79), 0.30 (95% CI −0.07 to 0.67), 0.22 (95% CI −0.07 to 0.51), 0.35 (95% CI −0.04 to 0.74), 0.25 (95% CI −0.07 to 0.56), respectively. This suggests that results of those between-groups analyses were not stable and reliable

Fig. 3

Sensitivity analysis of CSF Aβ42 levels in iNPH compared to healthy controls. Analyses were conducted on the raw data and the pooled estimate was −1.16 (95% CI −1.77 to −0.56). The direction and magnitude of pooled estimates did not change significantly after removing studies one by one, indicating that the results of the meta-analysis were relatively robust

Fig. 4

Sensitivity analysis of CSF t-tau levels in iNPH compared to AD. Analyses were conducted on the raw data and the pooled estimate was −1.29 (95% CI −1.99 to −0.58). The direction and magnitude of pooled estimates did not change significantly after removing studies one by one, indicating that the results of the meta-analysis were relatively robust

Fig. 5

Sensitivity analysis of CSF t-tau levels in iNPH compared to healthy controls. Analyses were conducted on the raw data and the pooled estimate was −0.81 (95% CI −1.53 to −0.10). The pooled estimate was different when the study of Agren-Wilsson et al. [12], Jeppsson et al. [25], Lim et al. [27] were omitted separately. The pooled estimates (95% CI) were −0.68 (95% CI −1.47 to 0.12), −0.59 (95% CI −1.29 to 0.11) and −0.83 (95% CI −1.68 to 0.21), respectively. This suggests that results of those between-groups analyses were not stable and reliable

Fig. 6

Sensitivity analysis of CSF p-tau levels in iNPH compared to AD. Analyses were conducted on the raw data and the pooled estimate was −1.57 (95% CI −2.39 to −0.76). The direction and magnitude of pooled estimates did not change significantly after removing studies one by one, indicating that the results of the meta-analysis were relatively robust

Fig. 7

Sensitivity analysis of CSF p-tau levels in iNPH compared to healthy controls. Analyses were conducted on the raw data and the pooled estimate was −1.14 (95% CI −1.40 to −0.87). The direction and magnitude of pooled estimates did not change significantly after removing studies one by one, indicating that the results of the meta-analysis were relatively robust