Key Points

Abstract

Introduction

Dementia with Lewy bodies (DLB) is the second most frequent neurodegenerative dementia after Alzheimer disease (AD).¹ The prognosis of DLB is usually poor,² even compared with AD; however, a large variation in the rate of decline is observed.³

The neuropathologic characteristics of DLB are heterogeneous. In addition to the defining α-synucleinopathy in the brainstem, limbic system, and cortical areas, there is often coexisting AD neuropathologic characteristics.⁴ Cerebrospinal fluid (CSF) concentrations of the core AD biomarkers (amyloid-β42/40 [Aβ42/40], total tau, and phosphorylated tau [p-tau]) are associated with molecular imaging of Aβ plaques and tau tangles and are usually within normal ranges in individuals with non-AD dementia.⁵ In DLB, however, about 50% of individuals have low CSF Aβ42 concentrations, which is indicative of amyloid plaque pathologic characteristics (ie, the key feature of AD neurodegeneration).⁵ The degree, type, location, and spread of neurodegenerative changes in individuals with DLB are highly variable, which may explain the heterogeneous clinical presentation and rate of cognitive decline observed in clinical studies.⁶ Imaging and CSF biomarker studies support the concept that there is an association between the presence of AD pathologic characteristics and the rate of cognitive decline in DLB.^7,8

There is now substantial evidence that blood biomarkers may be used to assess AD neuropathologic characteristics with high accuracy.^9,10 In this regard, p-tau species, including p-tau181 and p-tau231, have emerged as leading candidates, and the diagnostic and prognostic markers first shown in CSF samples have been widely replicated in blood samples.¹¹ Limited comparison data have suggested either a subtle difference or no difference between p-tau181 and p-tau231 levels in individuals with AD; however, these biomarkers have never been compared in individuals with DLB, to our knowledge.¹² In this study, we sought to assess plasma p-tau181 and p-tau231 concentrations and their association with cognition in a large international, multicenter DLB cohort from the European-DLB (E-DLB) Consortium.

Methods

A total of 1122 participants with plasma samples (with 371 participants with probable DLB overall) were available for this study were enrolled at 10 participating E-DLB centers with harmonized diagnostic procedures from January 1, 2002, to December 31, 2020.¹³ Participants were referred from outpatient clinics, including memory, movement disorders, geriatric medicine, psychiatric, and neurology clinics. Participants with acute delirium or terminal illness and patients with other previous major psychiatric or neurologic disorders were excluded, leaving a cohort of 987 clinically diagnosed participants with probable DLB (n=371), Parkinson disease (n=204), or AD (n=207) as well as healthy controls (HCs) (n=205). Additional information on the diagnostic examination program, inclusion and exclusion criteria, p-tau blood measurements, and statistical procedures for this study can be found in the eMethods in Supplement 1. Patients with AD and Parkinson disease, as well as HCs, were included for comparison (eTable 1 in Supplement 1). A subgroup of patients with AD (113 of 207 [54.6%]) had CSF confirmation based on reduced Aβ42 concentrations. Likewise, 34.0% of the DLB group (126 of 371) had Aβ42 concentrations obtained from CSF samples according to local reference values (eTable 2 in Supplement 1). Global cognitive function was assessed using the Mini-Mental State Examination (MMSE). A subgroup of 182 participants with probable DLB (eFigure 1 in Supplement 1) had annual longitudinal cognitive assessments for up to 5 years. Details of the sample collection procedures in each center are available in eTable 2 in Supplement 1. The local ethics committee at each center (eTable 1 in Supplement 1) approved the incorporation of data in this study and the analysis of samples abroad. The participants gave their written consent to use the deidentified results of their clinical, instrumental, and laboratory data for research purposes.

Plasma p-tau181 and p-tau231 concentrations were measured using a clinically validated, in-house single molecule array (Simoa) at the Clinical Neurochemistry Laboratory, Sahlgrenska University Hospital, Mölndal, Sweden (eTable 3 in Supplement 1).

Results

The baseline characteristics of the 987 participants (512 men [51.9%]; mean [SD] age, 70.0 [8.8] years) can be found in the Table. The DLB and AD groups did not differ significantly regarding age, sex, years of education, duration of symptoms, or MMSE baseline scores, but the DLB group was older than the HC group and included more men than did the AD and HC groups. In the DLB group with CSF analyses available, 63 of 126 patients (50.0%) had Aβ42 levels below the threshold.

Table.

Baseline Characteristics of the Cohort

Characteristic	Mean (SD) value					P value
	DLB (n=371 [37.6%])	AD (n=207 [20.9%])	PD (n=204 [20.7%])	HC (n=205 [20.8%])	Total (N=987)
Sex, No. (%)
Female	159 (42.9)	121 (58.5)	79 (38.7)	116 (56.6)	475 (48.1)	<.001
Male	212 (57.1)	86 (41.5)	125 (61.3)	89 (43.4)	512 (51.9)
Age, y	71.8 (8.0)	71.7 (8.0)	69.2 (9.9)	65.1 (8.0)	70.0 (8.8)	<.001a,b
Years of education	9.2 (4.2)	9.0 (4.0)	9.9 (4.7)	12.9 (3.6)	9.79 (4.5)	<.001a
Duration of symptoms, moc	25.2 (30.3)	25.6 (26.6)	71.5 (62.9)	53.6 (61.1)	41.2 (49.1)	<.001b
MMSE score	21.7 (6.0)	21.4 (5.3)	26.9 (3.7)	28.5 (1.7)	23.2 (5.8)	<.001a,b
Plasma p-tau231 level, pg/mL	12.8 (6.6)	16.9 (12.1)	12.3 (6.8)	11.0 (6.9)	13.2 (8.4)	<.001a,d
Plasma p-tau181 level, pg/mL	18.2 (9.5)	28.7 (28.4)	17.0 (8.5)	16.7 (18.5)	18.1 (17.2)	<.001a,d
Abnormal CSF Aβ42 level, No. (%)	63 (17.0)	113 (54.6)	NA	NA	NA	NA

Open in a separate window

Abbreviations: AD, Alzheimer disease; CSF, cerebrospinal fluid; DLB, dementia with Lewy bodies; HC, healthy control; MMSE, Mini-Mental State Examination; NA, not applicable; PD, Parkinson disease.

^a Differences between DLB and HC.

^b Differences between DLB and PD.

^c Duration from first motor or cognitive symptom.

^d Differences between DLB and AD.

Associations of Plasma p-tau Concentrations With Diagnoses

The plasma p-tau181 and p-tau231 concentrations in the different groups are shown in Figure 1. Patients with DLB had increased plasma p-tau181 and p-tau231 concentrations compared with the HC group (p-tau181, 2.58 pg/mL; 95% CI, 1.79-3.37 pg/mL; P=.007; p-tau231, 1.96 pg/mL; 95% CI, 1.37-2.56 pg/mL; P=.007) but lower p-tau181 and p-tau231 concentrations compared with the AD group (p-tau181, −3.53 pg/mL; 95% CI, −4.31 to −2.76 pg/mL; P<.001; p-tau231, −3.30 pg/mL; 95% CI, −3.98 to −2.71 pg/mL; P<.001). The numerical difference between DLB and AD increased when the analyses were performed using only the patients with AD (n=113) with confirmation of diagnosis based on low Aβ42 concentration in CSF samples. These results remained statistically significant after correcting for age and sex. When comparing patients with DLB and patients with AD with biomarker confirmation based on low Aβ42 concentrations in CSF samples, a greater effect size was seen for p-tau181 (Cohen d=0.53) compared with the full AD group (Cohen d=0.39). Effect sizes (Cohen d) for DLB compared with the HC group were similar for both plasma markers: 0.25 for p-tau231 and 0.23 for p-tau181. There were no statistically significant differences in either of the plasma p-tau concentrations between the Parkinson disease group and the DLB group. The DLB group with abnormal Aβ42 levels in CSF samples had higher plasma concentrations of both p-tau markers compared with those with normal levels in CSF samples (difference in p-tau181, −3.61 pg/mL; 95% CI, −5.43 to −1.79 pg/mL; P=.049; difference in p-tau231, −2.51 pg/mL; 95% CI, −3.63 to −1.39 pg/mL; P=.02) (Figure 1). The AUC values of the ROC curves indicating the overall biomarker performance across AD pathologic confirmed cases based on reduced Aβ42 levels in CSF samples obtained from patients with DLB were larger for p-tau181 than for p-tau231 (0.62 vs 0.56), but the change in the AUC of both biomarkers was not statistically significant (eFigure 2 in Supplement 1).

Open in a separate window

Figure 1.

Plasma p-tau181 and p-tau231 Concentrations

A, Distribution of plasma p-tau181 and p-tau231 concentrations across the different diagnostic groups. Dots indicate mean values, and the whiskers indicate their 95% CIs. Baseline concentrations of plasma p-tau181 and p-tau231 in patients with dementia with Lewy bodies (DLB) were significantly higher than those in the healthy control (HC) group but lower than those in the Alzheimer disease (AD) group. These differences between groups at baseline were evaluated using a Bonferroni post hoc test. B, Distribution of plasma p-tau181 and p-tau231 concentrations in patients with DLB with cerebrospinal fluid (CSF) amyloid-β42 (Aβ42) status. Baseline concentrations of plasma p-tau181 and p-tau231 in patients with DLB were significantly higher in the DLB group with abnormal CSF Aβ42 levels in both p-tau biomarkers compared with those with normal CSF Aβ42 levels. Patients were a mean (SD) age of 70.0 (8.8) years. PD indicates Parkinson disease.

Associations of Plasma p-tau Concentrations With Baseline and Longitudinal MMSE Score in Patients With DLB

We found that higher baseline plasma p-tau181 and p-tau231 concentrations were associated with a lower MMSE score at baseline (p-tau181, −0.092 MMSE points; 95% CI, −0.12 to −0.06 MMSE points; P=.001; p-tau231, −0.16 MMSE points; 95% CI, −0.21 to −0.12 MMSE points; P<.001) after adjusting for age and sex. After adjustment of the model for years of education, the p-tau231 concentration, but not the p-tau 181 concentration, remained significantly associated with a decrease in MMSE score (p-tau231, −0.049 MMSE points; 95% CI, −0.15 to −0.04 MMSE points; P=.049). The mean change estimated for the other diagnostic groups can be found in eTable 4 in Supplement 1. In the longitudinal model (DLB subgroup with longitudinal measurements of cognition [n=182] with a mean [SD] follow-up of 3.5 [1.7] years), the plasma p-tau181 concentration was associated with a decrease of −0.094 MMSE points per year (95% CI, −0.144 to −0.052 MMSE points; P=.02), whereas the plasma p-tau231 concentration was associated with an annual decrease of −0.130 MMSE points (95% CI, −0.201 to −0.071 MMSE points; P=.02) (Figure 2).

Open in a separate window

Figure 2.

Associations of Plasma p-tau181 and p-tau231 Concentration With Decrease in Mini-Mental State Examination (MMSE) Scores for Patients With Dementia With Lewy Bodies

A, MMSE score and plasma levels of p-tau181. B, MMSE score and plasma levels of p-tau231. The solid line shows the estimated marginal model for the MMSE scores for a typical patient at a mean (SD) age of 70.0 (8.8) years and 1.5 years of follow-up. The gray area represents the 95% CIs around the mean values. Both graphs show the decrease in MMSE scores as a function of the increase in plasma concentrations of p-tau181 and p-tau231.

Discussion

Here we examined the concentrations of plasma p-tau181 and p-tau231 in 987 patients from the E-DLB Consortium. We found that the baseline concentrations of plasma p-tau181 and p-tau231 in the DLB group were significantly higher than those in the HC group but lower than in the AD group. In addition, we found that both biomarkers were associated with more cognitive impairment in the DLB group at baseline as well as with more pronounced cognitive worsening over time during follow-up. Significance was lost after adjusting for years of education, possibly because of the smaller sample size with longitudinal follow-up and educational data available. Overall, the plasma p-tau181 concentration had a greater effect size when comparing DLB and AD, although it should not be used in isolation to make a clinical diagnosis. There was no difference between the use of p-tau181 and p-tau231 concentrations across confirmed AD pathologic characteristics based on reduced Aβ42 levels in CSF samples obtained from individuals with DLB. Moreover, plasma p-tau231 concentration had a larger association than p-tau181 concentration with cognition. The association with the decrease in MMSE score, although statistically significant, was small, and the clinical relevance may be questionable. Previously, plasma p-tau concentration has been shown to be strongly associated with cognitive decline in individuals with AD¹⁵ and is associated with AD neuropathologic characteristics. However, for individuals with DLB, plasma p-tau concentration has been examined only in small cohorts, to our knowledge.^9,16 Hall et al¹⁶ revealed that plasma p-tau concentration had a high correlation with abnormal tau and amyloid positron emission tomography scans from individuals with DLB, highlighting its usefulness as a marker for AD comorbid pathologic characteristics in DLB. Lantero Rodriguez et al,⁹ in neuropathologically confirmed cases, demonstrated increased plasma p-tau levels in those who had mixed AD and Lewy body pathologic characteristics a decade before death. In the small number of DLB cases without AD pathologic characteristics, the plasma p-tau levels were not elevated. In this study, we report that the levels of p-tau in individuals with DLB are intermediate between the HC and AD groups and significantly associated with cognitive decline in patients with DLB, suggesting that both p-tau181 and p-tau231 blood assays could be useful in assessing disease progression. Similar to Hall et al,¹⁶ we also demonstrate that plasma p-tau levels are elevated in DLB cases with AD comorbid pathologic findings. Further studies are needed to assess whether plasma p-tau levels can be used to assess the response of patients with DLB to treatments targeting AD pathophysiology.¹⁷

Conclusions

Plasma concentrations of p-tau181 and p-tau231 are elevated in patients with DLB and further elevated in patients with DLB with confirmed AD comorbid pathologic characteristics. Plasma p-tau represents a promising biomarker to identify AD pathologic characteristics and a faster cognitive decline in individuals with this neurologic disorder.

Notes

References