Key Points
Question
Can fluid biomarkers improve prediction of survival time in sporadic Creutzfeldt-Jakob disease (sCJD) above and beyond demographic and genetic biomarkers?
Findings
In this longitudinal cohort study including 188 participants with probable or definite sCJD and codon 129 genotyping, in addition to polymorphisms of prion protein gene codon 129 and baseline functional status, several cerebrospinal fluid–based and blood-based biomarkers were associated with survival in patients with sCJD. Total tau concentrations in the blood and cerebrospinal fluid appear to be the most promising.
Meaning
This study provides evidence that blood-based biomarkers can be used to predict survival in patients with sCJD, potentially improving clinical care and our ability to power treatment trials.
This longitudinal cohort study assesses whether plasma and cerebrospinal fluid biomarkers are associated with survival time in sporadic Creutzfeldt-Jakob disease.
Abstract
Importance
Fluid biomarkers that can predict survival time in sporadic Creutzfeldt-Jakob disease (sCJD) will be critical for clinical care and for treatment trials.
Objective
To assess whether plasma and cerebrospinal fluid (CSF) biomarkers are associated with survival time in patients with sCJD.
Design, Setting, and Participants
In this longitudinal cohort study, data from 193 patients with probable or definite sCJD who had codon 129 genotyping referred to a tertiary national referral service in the United States were collected from March 2004 to January 2018. Participants were evaluated until death or censored at the time of statistical analysis (range, 0.03-38.3 months). We fitted Cox proportional hazard models with time to event as the outcome. Fluid biomarkers were log-transformed, and models were run with and without nonfluid biomarkers of survival. Five patients were excluded because life-extending measures were performed.
Main Outcomes and Measures
Biomarkers of survival included sex, age, codon 129 genotype, Barthel Index, Medical Research Council Prion Disease Rating Scale, 8 CSF biomarkers (total tau [t-tau] level, phosphorylated tau [p-tau] level, t-tau:p-tau ratio, neurofilament light [NfL] level, β-amyloid 42 level, neuron-specific enolase level, 14-3-3 test result, and real-time quaking-induced conversion test), and 3 plasma biomarkers (t-tau level, NfL level, and glial fibrillary acidic protein level).
Results
Of the 188 included participants, 103 (54.8%) were male, and the mean (SD) age was 63.8 (9.2) years. Plasma t-tau levels (hazard ratio, 5.8; 95% CI, 2.3-14.8; P < .001) and CSF t-tau levels (hazard ratio, 1.6; 95% CI, 1.2-2.1; P < .001) were significantly associated with survival after controlling for codon 129 genotype and Barthel Index, which are also associated with survival time. Plasma and CSF t-tau levels were correlated (r = 0.74; 95% CI, 0.50-0.90; P < .001). Other fluid biomarkers associated with survival included plasma NfL levels, CSF NfL levels, t-tau:p-tau ratio, 14-3-3 test result, and neuron-specific enolase levels. In a restricted subset of 23 patients with data for all significant biomarkers, the hazard ratio for plasma t-tau level was more than 40% larger than any other biomarkers (hazard ratio, 3.4; 95% CI, 1.8-6.4).
Conclusions and Relevance
Cerebrospinal fluid and plasma tau levels, along with several other fluid biomarkers, were significantly associated with survival time in patients with sCJD. The correlation between CSF and plasma t-tau levels and the association of plasma t-tau level with survival time suggest that plasma t-tau level may be a minimally invasive fluid biomarker in sCJD that could improve clinical trial stratification and guide clinical care.
Introduction
Human prion disease, or Creutzfeldt-Jakob disease (CJD), is a rapidly progressive and fatal neurodegenerative condition caused by the propagation of abnormally folded prion proteins (PrPSc).1 Approximately 85% of human prion disease cases are sporadic CJD (sCJD),2,3 and therefore, sCJD has been the primary focus of clinical trials thus far.4 Although the typical duration of sCJD from time of diagnosis to death is 4 to 6 months on average, there is significant interindividual variability, with survival time ranging from weeks to several years,5,6,7,8 determined in part by prion protein gene (PRNP) codon 129 polymorphism and prion typing.9,10 Efforts are underway to bring preclinical therapeutics to human patients.11,12,13,14 As the pathogenic progression of protein aggregation in common neurodegenerative diseases (eg, Alzheimer disease) may occur via a prionlike propagation,15 CJD is considered to be an excellent model for treatments targeting the prionlike pathogenesis of neurodegeneration. Moreover, rapid decline in prion disease may allow clinical trials to detect a drug effect more quickly than trials enrolling patients with slowly progressing conditions.
Accurate prediction of disease duration has implications for clinical management, helping patients and families prepare for the disease course. Accurate prediction of survival time also affects clinical trial design, as stratifying patients based on expected disease course can improve a trial’s power,16 thereby reducing costs and ultimately maximizing the number of trials that can be funded. This is especially germane in sCJD, given the rarity of the disease and variable survival time.
Several studies have attempted to predict disease course in sCJD. Demographically, women and those with a younger age of onset may survive longer.6 A polymorphism at codon 129 of PRNP is one of the best-known predictors of clinical course in sCJD.6,17 Baseline cerebrospinal fluid (CSF) protein levels also have been studied predictors of survival time. Thus far, CSF total tau (t-tau) levels,18 the ratio of t-tau to phosphorylated tau (p-tau),19 neurofilament light chain (NfL) levels,20 and real-time quaking-induced conversion test (RT-QuIC) prion protein seeding activity21 have been reported to be associated with disease duration. More recently, blood-based biomarkers have been assessed, with one study showing increased serum t-tau and NfL concentrations in 33 patients with sCJD and 9 with genetic CJD compared with 40 healthy controls and 20 with other neurological disorders.22 In a study of 23 patients with sCJD,23 serum t-tau level was associated with rapidity of decline on the Medical Research Council (MRC) Prion Disease Rating Scale. Plasma biomarkers are desirable, given that quantifying proteins through a blood draw is less invasive than through CSF, which enables convenient clinical measurement.
The current study aims to replicate and build on previous studies of predictive fluid biomarkers in sCJD. We analyzed demographic characteristics, functional severity, and codon 129 genotype as predictors of survival in a large US cohort of patients with sCJD. We then investigated the association of a broad range of CSF and blood biomarkers with survival time and whether they offer unique predictive value after adjusting for known predictors.
Methods
Participants
Participants were referred to the University of California, San Francisco (UCSF) Memory and Aging Center from March 2004 to January 2018. Participants included patients with probable and definite (pathology-proven) sCJD with PRNP codon 129 polymorphism data available (performed at the National Prion Disease Pathology Surveillance Center, Cleveland, Ohio). A total of 193 research participants from the UCSF Memory and Aging Center provided written informed consent and underwent all or several of the following: cognitive testing, informant measures (eg, Barthel Index, MRC Scale [collected regularly starting May 2016]), standardized neurological examination, and CSF and blood sample collection. If fluid biomarkers were collected at multiple time points, we selected the first sample collected. Sample collection protocol and quantification of fluid biomarkers are presented in the eMethods in the Supplement. We excluded 5 participants who were placed on life-extending treatments, leaving 188 participants for survival analysis. Study partners provided written informed consent and assisted with providing history and answering informant questionnaires. This study was approved by the UCSF Internal Review Board.
Nonfluid Biomarkers of Survival Time
In the first set of survival analyses, 5 potential nonfluid biomarkers of survival (ie, sex, age, codon 129 genotype, Barthel Index, and MRC Scale) were fitted in 5 separate Cox proportional hazard models, with survival time represented in continuous months. In this first step of model building, variables that were significant were considered as covariates in subsequent analyses of fluid biomarkers. The MRC Scale was not considered as a covariate owing to insufficient sample size.
Fluid Biomarkers of Survival Time
First, all fluid biomarkers were fitted in separate Cox proportional hazard models that maximized the sample for each biomarker. Each model was fitted with and without the nonfluid biomarkers found to be significantly associated with survival in the first model-building step. Fluid biomarkers significantly associated with survival time were selected for a second set of analyses in which a restricted data set was created including only participants with data on all significant fluid biomarkers. We next evaluated all fluid biomarkers in separate Cox proportional hazard models in this restricted subset to assess whether all remained significantly associated with survival in this smaller sample.
Statistical Analysis
Length of survival was calculated from the date of the first lumbar puncture or blood draw at the UCSF Memory and Aging Center until death. Patients alive at analysis were censored at that date.
Fluid biomarker concentrations were log-transformed using the natural log to fulfill the normal distribution assumption and to enhance interpretation of the hazard ratios (HRs) across biomarkers. The RT-QuIC and 14-3-3 results were reported dichotomously and thus not log-transformed. All biomarkers with reported clinical cutoff values for prion disease were also analyzed as dichotomized predictors (eMethods in the Supplement). The CSF t-tau:p-tau ratio was calculated after log-transforming both variables. Hazard ratios for log-transformed biomarkers can be interpreted as the percentage increase in HR per log-unit change in the variable rather than per unstandardized unit change (eg, 1 pg/mL). Correlations between CSF and plasma NfL levels and CSF and plasma t-tau levels were conducted using a Pearson correlation on the log-transformed data, as were correlations among fluid biomarkers, age, and Barthel Index. All tests were 2-sided, and P values less than .05 were considered statistically significant. Statistical analyses were conducted in Stata version 14.2 (StataCorp), and plots were created in R version 3.5.2 (The R Foundation).
Results
Baseline Characteristics of the Sample
Baseline characteristics of the sample of 188 patients, including demographic characteristics, Mini-Mental Status Examination scores, functional ability, codon 129 genotype, and prion typing, when available, are shown in Table 1. Demographic characteristics for subgroups with biomarkers are shown in eTables 1-5 in the Supplement. Descriptive statistics for fluid biomarker values are presented in Table 2. The CSF t-tau, CSF and plasma NfL, and CSF NSE levels were the only biomarkers to show significant correlations with functional impairment level.
Table 1. Participant Demographic Characteristics.
| Characteristic | No. (%) |
|---|---|
| Total, No. | 188 |
| Age at study visit, y | |
| Mean (SD) | 63.8 (9.2) |
| Median (IQR) | 64 (57.5-70) |
| Range | 38.0-86.0 |
| Female | 85 (45.2) |
| Sporadic Creutzfeldt-Jakob disease | |
| Pathologically confirmed | 147 (78.2) |
| Probable | 41 (21.8) |
| PRNP codon 129 and molecular classificationa | 188 (100) |
| With typing | 134 (71.3) |
| MM | 80 (42.5) |
| MM 1 | 32 (23.9) |
| MM 2 | 17 (12.7) |
| MM 1 + 2 | 13 (9.7) |
| MV | 73 (38.8) |
| MV 1 | 17 (12.7) |
| MV 2 | 15 (11.2) |
| MV 1 + 2 | 18 (13.4) |
| VV | 35 (18.6) |
| VV 1 | 2 (1.5) |
| VV 2 | 15 (11.2) |
| VV 1 + 2 | 5 (3.7) |
| Barthel Index score at first visit | |
| Total patients with data | 100 (53.2) |
| Mean (SD) | 62.5 (33.8) |
| Median (IQR) | 70 (30.0-91.3) |
| Range | 0-100 |
| MRC Prion Disease Rating Scale score at first visit | |
| Total patients with data | 13 (6.9) |
| Mean (SD) | 15.6 (4.6) |
| Median (IQR) | 16 (13.5-18.5) |
| Range | 3-20 |
| Average time from first symptom to study visit, mob | |
| Total patients with data | 182 (96.8) |
| Mean (SD) | 8.4 (8.0) |
| Median (IQR) | 6.6 (2.7-11) |
| Range | 0.4-57.5 |
| Average time from first study visit to death, mo | |
| Total patients with data | 188 (100) |
| Mean (SD) | 6.7 (7.3) |
| Median (IQR) | 3.6 (1.3-9.8) |
| Range | 0.3-38.3 |
| Average time from first symptom to death, mo | |
| Total patients with data | 181 (96.3) |
| Mean (SD) | 14.8 (11.5) |
| Median (IQR) | 12.8 (5.9-20.6) |
| Range | 1.1-64.4 |
| Genotype, mean (SD) | |
| MM (n = 78) | 12.8 (10.5) |
| MV (n = 71) | 18.9 (12.7) |
| VV (n = 32) | 10.6 (7.5) |
Abbreviations: IQR, interquartile range; MRC, Medical Research Council.
Some participants were missing prion typing. Percentages for molecular classification only include those with prion typing.
For 5 participants, we could not confirm a reported date for first symptom. For all other patients, we used first symptoms even if they were not clearly neurological but rather neuropsychiatric (eg, clear behavior change), which may make our participants appear longer lived compared with some other studies that count from first obvious neurological symptom.
Table 2. Plasma and Cerebrospinal Fluid (CSF) Characteristics of Cohorta.
| Biomarker | Sample, Total (Died), No. | Positive/Total No. (%) | Median (IQR) [Range]b | Correlation With Age, rc | P Value | Correlation With Barthel Index, rc | P Value |
|---|---|---|---|---|---|---|---|
| Plasma | |||||||
| t-tau | 24 (24) | NA | 8 (5-15) [2-153] | 0.21 | .32 | −0.10 | .66 |
| NfL | 24 (24) | NA | 150 (63-296) [27-776] | 0.22 | .03 | −0.53 | .01 |
| GFAP | 24 (24) | NA | 695 (362-1095) [264-2139] | −0.15 | .48 | 0.27 | .23 |
| CSF | |||||||
| t-tau | 125 (122) | NA | 2179 (1074-6274) [111-18 895] | −0.02 | .80 | −0.31 | .003 |
| t-tau Cutoffd | 125 (122) | 89/125 (71) | NA | NA | NA | NA | NA |
| p-tau | 49 (48) | NA | 53 (39-70) [11-125] | 0.21 | .15 | −0.03 | .84 |
| t-tau:p-tau Ratio | 47 (46) | NA | 1.9 (1.8-2.3) [1.6-3.2] | 0.14 | .33 | 0.18 | .27 |
| NfL | 49 (48) | NA | 6400 (3599-11 757) [905-36 850] | −0.05 | .74 | −0.44 | .005 |
| Aβ42 | 49 (48) | NA | 391 (252-558) [98-1051] | 0.06 | .68 | 0.23 | .15 |
| NSE | 123 (120) | NA | 37 (22-63) [2-263] | 0.05 | .57 | −0.34 | <.001 |
| NSE cutoffd | 123 (120) | 79/123 (64) | NA | NA | NA | NA | NA |
| 14-3-3 Proteind | 162 (159) | 84/162 (52) | NA | NA | NA | NA | NA |
| Negative | NA | 30/162 (19) | NA | NA | NA | NA | NA |
| Inconclusive | NA | 48/162 (30) | NA | NA | NA | NA | NA |
| RT-QuICd | 64 (61) | 53/64 (83) | NA | NA | NA | NA | NA |
Abbreviations: Aβ42, β-amyloid 42; GFAP, glial fibrillary acidic protein; IQR, interquartile range; NA, not applicable; NfL, neurofilament light; NSE, neuron-specific enolase; p-tau, phosphorylated; RT-QuIC, real-time quaking-induced conversion test; t-tau, total tau.
All continuous biomarkers are reported as pg/mL, except for NSE, which is reported as ng/mL.
Median shown rather than mean owing to skewing.
Pearson correlations with log-transformed biomarkers.
Dichotomized, with positive being ≥1150 pg/mL (National Prion Disease Pathology Surveillance Center) or ≥1200 pg/mL (Athena Diagnostics).
Correlations of Plasma t-tau, Plasma NfL, CSF t-tau, and CSF NfL Levels
Strong correlations were observed between plasma and CSF NfL concentrations (r = 0.80; 95% CI, 0.60-0.91; P < .001) and between plasma and CSF t-tau concentrations (r = 0.74; 95% CI, 0.48-0.88; P < .001). Cerebrospinal fluid t-tau and CSF NfL concentrations were also correlated (r = 0.37; 95% CI, 0.09-0.59; P = .01). There was a positive but nonsignificant correlation between plasma NfL and t-tau concentrations (r = 0.38; 95% CI, −0.01 to 0.68; P = .06).
Biomarkers Associated With Survival
We first evaluated the association of age, sex, codon 129 genotype, MRC Scale, and Barthel Index with survival. The Barthel Index, MRC Scale, and codon 129 genotype, but not age or sex, were significantly associated with survival time (Table 2). For codon 129 genotypes, VV had the highest HR, followed by MM and then MV; the only between-group difference reaching statistical significance was VV compared with MV (HR, 1.61; 95% CI, 1.07-2.43; P = .02). Lower baseline levels of function (measured by Barthel Index or MRC Scale) predicted a faster disease course. Based on our model-building criteria, in subsequent analyses, we fitted Cox models with and without controlling for Barthel Index and codon 129 genotype. Interestingly, when both Barthel Index and codon 129 genotype were entered simultaneously as predictors of survival (n = 102; 2 censored), only Barthel Index remained statistically significant (HR, 0.99; 95% CI, 0.98-0.995; P = .001).
Each fluid biomarker was assessed as a predictor of survival time (Table 3). Greater baseline levels of plasma t-tau and NfL levels were associated with shorter survival, and the association of plasma t-tau level with survival time remained after controlling for Barthel Index and codon 129 genotype. Importantly, plasma t-tau level and Barthel Index (HR, 0.98; 95% CI, 0.96-0.99; P = .008) each had independent value in predicting survival. The HRs for all CSF biomarkers were in the expected direction (higher levels associated with shorter survival), with t-tau level, t-tau:p-tau ratio, NfL level, and NSE level reaching statistical significance. Positive results for 14-3-3 protein and based on cutoff values for NSE and t-tau levels were associated with a shorter time until death. Similar to the results for plasma biomarkers, CSF t-tau level remained associated with survival after controlling for Barthel Index and codon 129 genotype, as did CSF t-tau:p-tau ratio, NSE level, and 14-3-3 result. The association of plasma and CSF t-tau levels with survival are displayed using Kaplan-Meier curves in the Figure.
Table 3. Cox Proportional Hazard Models With Fluid Biomarkers Associated With Survival.
| Biomarker | No Covariates | Covariates: Barthel Index and Codon 129 Genotype | ||||
|---|---|---|---|---|---|---|
| Sample, Total (Died), No. | HR (95% CI) | P Value | Sample, Total (Died), No. | HR (95% CI) | P Value | |
| Covariate | ||||||
| Age | 188 (186) | 1.01 (0.99-1.03) | .29 | NA | NA | NA |
| Sex | 188 (186) | 1.08 (0.81-1.45) | .60 | NA | NA | NA |
| Codon 129 genotype | 188 (186) | 1.24 (1.02-1.51) | .03 | NA | NA | NA |
| Barthel Indexa | 155 (152) | 0.98 (0.98-0.99) | <.001 | NA | NA | NA |
| MRC Prion Disease Rating Scale | 24 (21) | .90 (0.82-0.98) | .03 | NA | NA | NA |
| Plasma | ||||||
| t-tau | 24 (24) | 3.37 (1.83-6.19) | <.001 | 21 (21) | 5.77 (2.26-14.75) | <.001 |
| NfL | 24 (24) | 2.08 (1.22-3.54) | .007 | 21 (21) | 1.95 (0.92-4.10) | .08 |
| GFAP | 24 (24) | 0.76 (0.37-1.57) | .46 | 21 (21) | 0.61 (0.21-1.73) | .35 |
| CSF | ||||||
| t-tau | 125 (122) | 1.60 (1.34-1.92) | <.001 | 83 (81) | 1.60 (1.22-2.10) | .001 |
| t-tau Cutoffb | 125 (122) | 1.90 (1.27-2.84) | .002 | 83 (81) | 1.17 (0.71-1.95) | .53 |
| p-tau | 49 (48) | 1.10 (0.63-1.93) | .74 | 40 (39) | 0.81 (0.42-1.56) | .53 |
| t-tau:p-tau Ratio | 47 (46) | 2.14 (1.09-4.21) | .03 | 39 (38) | 5.13 (1.39-18.97) | .01 |
| NfL | 49 (48) | 1.61 (1.11-2.35) | .01 | 40 (39) | 1.32 (0.85-2.03) | .22 |
| Aβ42 | 49 (48) | 0.65 (0.33-1.27) | .21 | 40 (39) | 0.52 (0.24-1.16) | .11 |
| NSE | 123 (120) | 1.46 (1.17-1.82) | .001 | 84 (82) | 1.62 (1.06-2.48) | .03 |
| NSE cutoffb | 123 (120) | 1.50 (1.03-2.18) | .04 | 84 (82) | 1.15 (0.68-1.93) | .61 |
| 14-3-3 Protein | 162 (159) | 1.75 (1.27-2.40) | .001 | 89 (87) | 1.78 (1.11-2.83) | .02 |
| RT-QuICb | 64 (61) | 1.81 (0.91-3.60) | .09 | 40 (38) | 1.92 (0.83-4.43) | .13 |
Abbreviations: Aβ42, β-amyloid 42; GFAP, glial fibrillary acidic protein; HR, hazard ratio; IQR, interquartile range; MRC, Medical Research Council; NA, not applicable; NfL, neurofilament light; NSE, neuron-specific enolase; p-tau, phosphorylated; RT-QuIC, real-time quaking-induced conversion test; t-tau, total tau.
Barthel Index score was divided by 5 to put it on same scale as the MRC Scale.
Dichotomized and therefore not log-transformed.
Figure. Kaplan-Meier Survival Curves for Plasma and Cerebrospinal Fluid (CSF) Total Tau Levels.
Kaplan-Meier curves displaying time from blood draw (A) or lumbar puncture (B) to death in sporadic Creutzfeldt-Jakob disease as a function of total tau levels. A, A median split was conducted on plasma t-tau level (8.3 pg/mL) for illustrative purposes. Higher concentrations at baseline were associated with shorter survival time. All patients were deceased at the time of analysis. B, Because the sample size was larger, CSF total tau concentration was binned in tertiles. Censored participants are denoted by hash marks.
We next evaluated the 7 fluid biomarkers that reached statistical significance in a restricted subsample of 23 patients with data for all of these biomarkers (Table 4). There were no significant differences between this subsample and the larger sample of 188 patients in age, Barthel Index, time from lumbar puncture to event (death or censoring), or total disease duration. All fluid biomarkers remained significantly associated with survival, with the exception of t-tau:p-tau ratio and 14-3-3. Compared with CSF t-tau level, the HR for plasma t-tau level (HR, 3.41; 95% CI, 1.82-6.39; P < .001) was nearly twice as large and was more than 40% larger than the next highest HR (plasma NfL level).
Table 4. Hazard Ratios (HRs) for Biomarkers in the Restricted Data Set in Which Patients Have Values for All Biomarkersa.
| Biomarkerb | HR (95% CI) | P Value |
|---|---|---|
| Plasma t-tau | 3.41 (1.82-6.39) | <.001 |
| Plasma NfL | 1.97 (1.17-3.30) | .01 |
| CSF NfL | 1.91 (1.08-3.38) | .03 |
| CSF t-tau | 1.80 (1.08-3.01) | .03 |
| t-tau:p-tau Ratio | 1.81 (0.78-4.15) | .16 |
| CSF 14-3-3 protein | 1.75 (0.74-4.16) | .20 |
| CSF NSE | 2.33 (1.11-4.92) | .03 |
Abbreviations: CSF, cerebrospinal fluid; NfL, neurofilament light; NSE, neuron-specific enolase; p-tau, phosphorylated; t-tau, total tau.
All 23 patients included in this analysis were deceased at the time of analysis.
All biomarkers have been log-transformed to enhance comparability and achieve normality.
Discussion
Our results support the potential value of several fluid biomarkers as prognostic tools in sCJD. Total tau concentration, measured in either blood or CSF, was strongly associated with survival time. Although NfL and NSE levels and 14-3-3 results were also associated with survival, the HR associated with plasma t-tau level was more than 40% higher than other fluid biomarkers of interest. These findings further bolster the value of blood-based biomarkers based on their minimally invasive and relatively inexpensive nature and build on prior studies that suggested patients with sCJD and controls can be discriminated with relatively high accuracy using blood-based assays.23,24 Our results are generally consistent with previous work by Thompson et al23 showing that plasma t-tau level correlates with the rate of disease progression in sCJD. They did not find an association of plasma t-tau level with survival time, although they did not control for baseline functional severity. Nevertheless, we found plasma t-tau level correlated with survival time irrespective of functional severity. We also showed that plasma t-tau level showed greater predictive value than several other plasma and CSF biomarkers, including plasma glial fibrillary acidic protein, plasma and CSF NfL, and CSF t-tau concentrations, even after controlling for codon 129 genotype and baseline functioning. Furthermore, worse baseline functional status (ie, Barthel Index or MRC Scale) was associated with shorter time until death. Importantly, when baseline functional status and plasma t-tau levels were modeled together, both were strong, independent predictors of survival time. This suggests that clinical measures and plasma t-tau level could be combined to further improve prediction accuracy.
Tau, a microtubule-stabilizing protein found in central nervous system neurons and glial cells, is released extracellularly during neuronal injury and death, and it is elevated in the CSF both in patients with acute brain injuries, such as hypoxia25 and head trauma,26 and in several chronic neurodegenerative diseases.27,28 Hyperphosphorylated tau is not uncommonly found in sCJD on neuropathology (eMethods in the Supplement).29 Although CSF t-tau concentration has long been known as a potential diagnostic biomarker in sCJD30,31—and a few studies have implicated it as a predictive biomarker18,19—plasma t-tau level is a relatively new biomarker. Plasma tau level is currently quantified using research-use-only commercial assays, but efforts are underway to develop these into assays validated for use in clinical diagnostics. One of the first, albeit small, studies of serum t-tau levels found greater elevations in sCJD (n = 12) compared with other neurodegenerative conditions (n = 19), including some nonprion rapidly progressive dementia cases.32 These findings have since been confirmed in 2 larger cohorts,23,24 although only the study by Thompson et al23 assessed its validity as a predictive biomarker. Although elevated levels of plasma t-tau also have been reported in frontotemporal dementia33 and Alzheimer disease34 and are correlated with cognitive decline in Alzheimer disease,35 concerns have been raised about the potential of plasma t-tau level as a biomarker36 because it may be unstable or rapidly cleared in blood, as it is not correlated with CSF levels in Alzheimer disease.34,37,38 In our study, however, plasma and CSF t-tau levels were highly correlated (r = 0.74; 95% CI, 0.48-0.88; P < .001), higher in magnitude than previously reported associations (ie, r = 0.59; P < .001).24 One explanation for the high correlation between CSF and plasma tau level in sCJD is that the rapidity and amount of neurodegeneration may result in continuous, high quantities of tau in the periphery that overcome the peripheral mechanisms of elimination. Others have argued that higher plasma t-tau concentrations in sCJD may be driven by preterminal cases having higher t-tau levels.23 Our results, however, indicate that high levels of plasma tau may be a biomarker of disease intensity/rapidity regardless of the clinical severity. The correlation between CSF and plasma t-tau level, together with the concordance of CSF and plasma t-tau level in predicting survival time, further strengthens our finding that plasma tau level may be a useful biomarker in this cohort.
In addition to t-tau level, other biomarkers were promising predictors of survival. Plasma NfL level and CSF t-tau level, t-tau:p-tau ratio, NfL level, 14-3-3 result, and NSE level (continuously or dichotomously) were all significantly associated with survival time, albeit less than plasma t-tau level. Four of these biomarkers—CSF t-tau level, t-tau:p-tau ratio, 14-3-3 result, and NSE level (continuous)—were shown to provide additional predictive value of survival above and beyond codon 129 genotype and Barthel Index. These results confirm prior findings by other groups that suggest CSF t-tau level18 and NfL level20 may inform prediction of survival time in CJD. Prior studies have shown the t-tau:p-tau ratio to help in sCJD diagnosis,19,39 and one of these studies found that it predicted survival time.19 We extend these results by controlling for functional status and codon 129 genotype and show that CSF NfL level no longer remains significant after controlling for these biomarkers. Although very promising for diagnosis, RT-QuIC does not appear to be associated with survival in its current form as a binary (ie, positive or negative) variable. It is worth noting, however, that it is now possible to quantify seeding activity as a continuous variable,21 which may have additive value in forecasting disease course. β-Amyloid 42 levels in the CSF were not associated with survival.
Previous data showing the role of a common codon 129 polymorphism in PRNP as a modifier of disease trajectory in sCJD were replicated,6,17 consistent with experimental literature indicating that alterations of PRNP can affect pathogenesis.40,41 We found that VV had the highest HR, followed by MM and then MV, confirming prior studies showing that patients with MV have the slowest disease progression.6,17 In 2016, Mead et al17 reported codon 129 genotype as a predictor of disease progression, using a model-derived estimate of the rate of decline on the MRC Scale as an outcome among 139 patients with sCJD and 15 patients with fast-progressing genetic prion disease, finding that MM, rather than VV, as in our study, predicted the fastest rate of clinical decline. Despite the difference in our findings of MM vs VV (which might be because of their inclusion of patients with fast-progressing genetic prion disease or other factors), the results of Mead et al17 taken together with our larger study strongly suggest that codon 129 polymorphism should be considered as a factor by which to stratify treatment trials. Whereas sex and age were not significant predictors in our cohort, a very large study by Pocchiari et al6 including 2034 patients with sCJD found younger age and female sex as well as codon 129 heterozygosity (n = 1453) to be associated with longer survival time.
Limitations
Although very promising, our results are not without limitations. There was only a relatively small subsample of patients with all plasma and CSF biomarkers. As we have previously reported, disease duration or survival of our sCJD cohort is somewhat longer than some other literature, possibly because of our extensive methods for identifying the first symptom.31 These results require independent validation in a larger cohort, controlling for other potential predictors of decline, such as baseline functional status, codon 129 genotype, age, and sex. The sample with MRC Scale data in our study was relatively small because we did not begin collecting these data until recently, but nevertheless, this scale appears to be a promising predictor of decline. Another limitation is that the plasma biomarkers in this study, one of which showed great promise for predicting survival, were assayed using a research protocol. Widespread clinical use of these biomarkers will require well-validated commercial assays, development of which are underway. Similar to the effort in the field of Alzheimer disease,24,42 it will be important to establish a consensus on the procedures and platforms used to calculate these concentrations. Future work replicating these findings and optimizing cutoffs will be required before moving these biomarkers into the clinic or using them for clinical trials. Plasma tau level was not analyzed as a diagnostic biomarker in this study; future work should explore this potential. Because of the comprehensive nature of this analysis, we chose not to examine neuroimaging predictors of survival. This is an important topic that is being addressed in ongoing studies.
Conclusions
In conclusion, we provide initial evidence that plasma t-tau concentration in patients with sCJD is a significant predictor of survival beyond functional status and PRNP codon 129 polymorphism. Furthermore, we extend prior studies by analyzing both plasma and CSF t-tau levels in the same sample and showing that plasma t-tau level is more strongly associated with survival than CSF t-tau levels. These results are particularly encouraging because blood-based biomarkers are relatively less invasive to obtain than CSF biomarkers, which require a lumbar puncture. This study adds to research in other dementia disorders, such as Alzheimer disease,34,36,43 suggesting that blood-based assays may revolutionize the way that neurodegenerative diseases are diagnosed and monitored. We also provide, to our knowledge, the first evidence that plasma glial fibrillary acidic protein and CSF β-amyloid 42 concentrations do not appear to have value as prognostic biomarkers for survival time, although they could still be promising for differential diagnosis. Plasma t-tau level as well as some CSF biomarkers, functional status, and codon 129 polymorphisms are promising markers of survival time in sCJD and should be considered when designing clinical trials.
eMethods.
eTable 1. Participant demographic characteristics for patients with CSF t-tau measurements.
eTable 2. Participant demographic characteristics for patients with CSF NSE measurements.
eTable 3. Participant demographic characteristics for patients with RT-QuIC measurements.
eTable 4. Participant demographic characteristics for patients with plasma measurements.
eReferences.
eTable 5. Participant demographic characteristics for patients with CSF measurements (analyzed in Sweden).
References
- 1.Takada LT, Geschwind MD. Prion diseases. Semin Neurol. 2013;33(4):348-356. doi: 10.1055/s-0033-1359314 [DOI] [PubMed] [Google Scholar]
- 2.Ladogana A, Puopolo M, Croes EA, et al. . Mortality from Creutzfeldt-Jakob disease and related disorders in Europe, Australia, and Canada. Neurology. 2005;64(9):1586-1591. doi: 10.1212/01.WNL.0000160117.56690.B2 [DOI] [PubMed] [Google Scholar]
- 3.Masters CL, Harris JO, Gajdusek DC, Gibbs CJ Jr, Bernoulli C, Asher DM. Creutzfeldt-Jakob disease: patterns of worldwide occurrence and the significance of familial and sporadic clustering. Ann Neurol. 1979;5(2):177-188. doi: 10.1002/ana.410050212 [DOI] [PubMed] [Google Scholar]
- 4.Mead S, Tagliavini F. Clinical trials. Handb Clin Neurol. 2018;153:431-444. doi: 10.1016/B978-0-444-63945-5.00024-6 [DOI] [PubMed] [Google Scholar]
- 5.Johnson DY, Dunkelberger DL, Henry M, et al. . Sporadic Jakob-Creutzfeldt disease presenting as primary progressive aphasia. JAMA Neurol. 2013;70(2):254-257. doi: 10.1001/2013.jamaneurol.139 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Pocchiari M, Puopolo M, Croes EA, et al. . Predictors of survival in sporadic Creutzfeldt-Jakob disease and other human transmissible spongiform encephalopathies. Brain. 2004;127(pt 10):2348-2359. doi: 10.1093/brain/awh249 [DOI] [PubMed] [Google Scholar]
- 7.Puoti G, Bizzi A, Forloni G, Safar JG, Tagliavini F, Gambetti P. Sporadic human prion diseases: molecular insights and diagnosis. Lancet Neurol. 2012;11(7):618-628. doi: 10.1016/S1474-4422(12)70063-7 [DOI] [PubMed] [Google Scholar]
- 8.Brown P, Cathala F, Castaigne P, Gajdusek DC. Creutzfeldt-Jakob disease: clinical analysis of a consecutive series of 230 neuropathologically verified cases. Ann Neurol. 1986;20(5):597-602. doi: 10.1002/ana.410200507 [DOI] [PubMed] [Google Scholar]
- 9.Collins SJ, Sanchez-Juan P, Masters CL, et al. . Determinants of diagnostic investigation sensitivities across the clinical spectrum of sporadic Creutzfeldt-Jakob disease. Brain. 2006;129(pt 9):2278-2287. doi: 10.1093/brain/awl159 [DOI] [PubMed] [Google Scholar]
- 10.Parchi P, Giese A, Capellari S, et al. . Classification of sporadic Creutzfeldt-Jakob disease based on molecular and phenotypic analysis of 300 subjects. Ann Neurol. 1999;46(2):224-233. doi: [DOI] [PubMed] [Google Scholar]
- 11.White AR, Enever P, Tayebi M, et al. . Monoclonal antibodies inhibit prion replication and delay the development of prion disease. Nature. 2003;422(6927):80-83. doi: 10.1038/nature01457 [DOI] [PubMed] [Google Scholar]
- 12.Trevitt CR, Collinge J. A systematic review of prion therapeutics in experimental models. Brain. 2006;129(pt 9):2241-2265. doi: 10.1093/brain/awl150 [DOI] [PubMed] [Google Scholar]
- 13.Wagner J, Ryazanov S, Leonov A, et al. . Anle138b: a novel oligomer modulator for disease-modifying therapy of neurodegenerative diseases such as prion and Parkinson’s disease. Acta Neuropathol. 2013;125(6):795-813. doi: 10.1007/s00401-013-1114-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Silber BM, Rao S, Fife KL, et al. . Pharmacokinetics and metabolism of 2-aminothiazoles with antiprion activity in mice. Pharm Res. 2013;30(4):932-950. doi: 10.1007/s11095-012-0912-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jucker M, Walker LC. Self-propagation of pathogenic protein aggregates in neurodegenerative diseases. Nature. 2013;501(7465):45-51. doi: 10.1038/nature12481 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tsai RM, Boxer AL. Therapy and clinical trials in frontotemporal dementia: past, present, and future. J Neurochem. 2016;138(suppl 1):211-221. doi: 10.1111/jnc.13640 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Mead S, Burnell M, Lowe J, et al. . Clinical trial simulations based on genetic stratification and the natural history of a functional outcome measure in Creutzfeldt-Jakob disease. JAMA Neurol. 2016;73(4):447-455. doi: 10.1001/jamaneurol.2015.4885 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Karch A, Hermann P, Ponto C, et al. . Cerebrospinal fluid tau levels are a marker for molecular subtype in sporadic Creutzfeldt-Jakob disease. Neurobiol Aging. 2015;36(5):1964-1968. doi: 10.1016/j.neurobiolaging.2015.01.021 [DOI] [PubMed] [Google Scholar]
- 19.Skillbäck T, Rosén C, Asztely F, Mattsson N, Blennow K, Zetterberg H. Diagnostic performance of cerebrospinal fluid total tau and phosphorylated tau in Creutzfeldt-Jakob disease: results from the Swedish Mortality Registry. JAMA Neurol. 2014;71(4):476-483. doi: 10.1001/jamaneurol.2013.6455 [DOI] [PubMed] [Google Scholar]
- 20.Zerr I, Schmitz M, Karch A, et al. . Cerebrospinal fluid neurofilament light levels in neurodegenerative dementia: evaluation of diagnostic accuracy in the differential diagnosis of prion diseases. Alzheimers Dement. 2018;14(6):751-763. doi: 10.1016/j.jalz.2017.12.008 [DOI] [PubMed] [Google Scholar]
- 21.Foutz A, Appleby BS, Hamlin C, et al. . Diagnostic and prognostic value of human prion detection in cerebrospinal fluid. Ann Neurol. 2017;81(1):79-92. doi: 10.1002/ana.24833 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Steinacker P, Blennow K, Halbgebauer S, et al. . Neurofilaments in blood and CSF for diagnosis and prediction of onset in Creutzfeldt-Jakob disease. Sci Rep. 2016;6(1):38737. doi: 10.1038/srep38737 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Thompson AGB, Luk C, Heslegrave AJ, et al. . Neurofilament light chain and tau concentrations are markedly increased in the serum of patients with sporadic Creutzfeldt-Jakob disease, and tau correlates with rate of disease progression. J Neurol Neurosurg Psychiatry. 2018;89(9):955-961. doi: 10.1136/jnnp-2017-317793 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kovacs GG, Andreasson U, Liman V, et al. . Plasma and cerebrospinal fluid tau and neurofilament concentrations in rapidly progressive neurological syndromes: a neuropathology-based cohort. Eur J Neurol. 2017;24(11):1326-e77. doi: 10.1111/ene.13389 [DOI] [PubMed] [Google Scholar]
- 25.Randall J, Mörtberg E, Provuncher GK, et al. . Tau proteins in serum predict neurological outcome after hypoxic brain injury from cardiac arrest: results of a pilot study. Resuscitation. 2013;84(3):351-356. doi: 10.1016/j.resuscitation.2012.07.027 [DOI] [PubMed] [Google Scholar]
- 26.Franz G, Beer R, Kampfl A, et al. . Amyloid beta 1-42 and tau in cerebrospinal fluid after severe traumatic brain injury. Neurology. 2003;60(9):1457-1461. doi: 10.1212/01.WNL.0000063313.57292.00 [DOI] [PubMed] [Google Scholar]
- 27.Zetterberg H. Review: tau in biofluids: relation to pathology, imaging and clinical features. Neuropathol Appl Neurobiol. 2017;43(3):194-199. doi: 10.1111/nan.12378 [DOI] [PubMed] [Google Scholar]
- 28.Mattsson N, Zetterberg H, Bianconi S, et al. . Gamma-secretase-dependent amyloid-beta is increased in Niemann-Pick type C: a cross-sectional study. Neurology. 2011;76(4):366-372. doi: 10.1212/WNL.0b013e318208f4ab [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Tousseyn T, Bajsarowicz K, Sánchez H, et al. . Prion disease induces Alzheimer disease-like neuropathologic changes. J Neuropathol Exp Neurol. 2015;74(9):873-888. doi: 10.1097/NEN.0000000000000228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Chohan G, Pennington C, Mackenzie JM, et al. . The role of cerebrospinal fluid 14-3-3 and other proteins in the diagnosis of sporadic Creutzfeldt-Jakob disease in the UK: a 10-year review. J Neurol Neurosurg Psychiatry. 2010;81(11):1243-1248. doi: 10.1136/jnnp.2009.197962 [DOI] [PubMed] [Google Scholar]
- 31.Forner SA, Takada LT, Bettcher BM, et al. . Comparing CSF biomarkers and brain MRI in the diagnosis of sporadic Creutzfeldt-Jakob disease. Neurol Clin Pract. 2015;5(2):116-125. doi: 10.1212/CPJ.0000000000000111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Noguchi-Shinohara M, Hamaguchi T, Nozaki I, Sakai K, Yamada M. Serum tau protein as a marker for the diagnosis of Creutzfeldt-Jakob disease. J Neurol. 2011;258(8):1464-1468. doi: 10.1007/s00415-011-5960-x [DOI] [PubMed] [Google Scholar]
- 33.Foiani MS, Woollacott IO, Heller C, et al. . Plasma tau is increased in frontotemporal dementia. J Neurol Neurosurg Psychiatry. 2018;89(8):804-807. doi: 10.1136/jnnp-2017-317260 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Mattsson N, Zetterberg H, Janelidze S, et al. ; ADNI Investigators . Plasma tau in Alzheimer disease. Neurology. 2016;87(17):1827-1835. doi: 10.1212/WNL.0000000000003246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Mielke MM, Hagen CE, Wennberg AMV, et al. . Association of plasma total tau level with cognitive decline and risk of mild cognitive impairment or dementia in the Mayo Clinic Study on Aging. JAMA Neurol. 2017;74(9):1073-1080. doi: 10.1001/jamaneurol.2017.1359 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Zetterberg H, Blennow K. From cerebrospinal fluid to blood: the third wave of fluid biomarkers for Alzheimer’s disease. J Alzheimers Dis. 2018;64(s1):S271-S279. doi: 10.3233/JAD-179926 [DOI] [PubMed] [Google Scholar]
- 37.Zetterberg H, Wilson D, Andreasson U, et al. . Plasma tau levels in Alzheimer’s disease. Alzheimers Res Ther. 2013;5(2):9. doi: 10.1186/alzrt163 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Zetterberg H, Blennow K. Chronic traumatic encephalopathy: fluid biomarkers. Handb Clin Neurol. 2018;158:323-333. doi: 10.1016/B978-0-444-63954-7.00030-6 [DOI] [PubMed] [Google Scholar]
- 39.Dorey A, Tholance Y, Vighetto A, et al. . Association of cerebrospinal fluid prion protein levels and the distinction between Alzheimer disease and Creutzfeldt-Jakob disease. JAMA Neurol. 2015;72(3):267-275. doi: 10.1001/jamaneurol.2014.4068 [DOI] [PubMed] [Google Scholar]
- 40.Palmer MS, Dryden AJ, Hughes JT, Collinge J. Homozygous prion protein genotype predisposes to sporadic Creutzfeldt-Jakob disease. Nature. 1991;352(6333):340-342. doi: 10.1038/352340a0 [DOI] [PubMed] [Google Scholar]
- 41.Collinge J, Clarke AR. A general model of prion strains and their pathogenicity. Science. 2007;318(5852):930-936. doi: 10.1126/science.1138718 [DOI] [PubMed] [Google Scholar]
- 42.Shaw LM, Arias J, Blennow K, et al. . Appropriate use criteria for lumbar puncture and cerebrospinal fluid testing in the diagnosis of Alzheimer’s disease. Alzheimers Dement. 2018;14(11):1505-1521. doi: 10.1016/j.jalz.2018.07.220 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Mattsson N, Andreasson U, Zetterberg H, Blennow K; Alzheimer’s Disease Neuroimaging Initiative . Association of plasma neurofilament light with neurodegeneration in patients with Alzheimer disease. JAMA Neurol. 2017;74(5):557-566. doi: 10.1001/jamaneurol.2016.6117 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods.
eTable 1. Participant demographic characteristics for patients with CSF t-tau measurements.
eTable 2. Participant demographic characteristics for patients with CSF NSE measurements.
eTable 3. Participant demographic characteristics for patients with RT-QuIC measurements.
eTable 4. Participant demographic characteristics for patients with plasma measurements.
eReferences.
eTable 5. Participant demographic characteristics for patients with CSF measurements (analyzed in Sweden).