Real-Time Quaking-Induced Conversion Assays for Prions: Applying a Sensitive but Imperfect Test in Clinical Practice

Abstract

Background:

Real-time quaking-induced conversion (RT-QuIC) assays offer a sensitive and specific means for detection of prions, although false negative results are recognized in clinical practice. We profile the clinical, laboratory, and pathologic features associated with false negative RT-QuIC assays and extend these to frame the diagnostic approach to patients with suspected prion disease.

Methods:

113 patients with probable or definite prion disease were assessed at Mayo Clinic (Rochester, MN; Jacksonville, FL; Scottsdale, AZ) or Washington University School of Medicine (St. Louis, MO) from 2013–2021. RT-QuIC testing for prions was performed in CSF at the National Prion Disease Pathology Surveillance Center (Cleveland, OH).

Results:

Initial RT-QuIC testing was negative in 13/113 patients (sensitivity 88.5%). RT-QuIC negative patients were younger (median 52.0 years vs 66.1 years, p<0.001). Other demographic and presenting features, and CSF cell count, protein, and glucose levels were similar in RT-QuIC negative and positive patients. Frequency of 14–3-3 positivity (4/13 vs 77/94, p<0.001) and median CSF total tau levels were lower in RT-QuIC negative patients (2517 vs 4001 pg/mL, p=0.020), while time from symptom onset to first presentation (153 vs 47 days, p=0.001) and symptomatic duration (710 vs 148 days, p=0.001) were longer.

Conclusions:

RT-QuIC is a sensitive yet imperfect measure necessitating incorporation of other test results when evaluating patients with suspected prion disease. Patients with negative RT-QuIC had lower markers of neuronal damage (CSF total-tau and protein 14–3-3) and longer symptomatic duration of disease suggesting that false negative RT-QuIC testing associates with a more indolent course.

Introduction

Creutzfeldt-Jakob disease (CJD) is a rapidly progressive and universally fatal transmissible spongiform encephalopathy resulting from the propagation of misfolded prion proteins in the brain. Most cases develop sporadically (sCJD), presenting with rapidly progressive dementia, neuropsychiatric symptoms, ataxia, or movement disorders; median survival is less than 12 months.^{1, 2} The heterogeneous symptomatology, overlap with other causes of dementia (including treatment-responsive mimics^{27, 28, 29}), and potential for iatrogenic transmission³ exemplify the need for an accurate pre-mortem diagnosis that is supported by sensitive, specific, and accessible diagnostic tests.³

The development and refinement of real-time quaking-induced conversion (RT-QuIC) measures capable of detecting small amounts of abnormal prions in CSF and other biospecimens (e.g., nasal brushings^{4, 5}) has revolutionized the diagnostic approach to CJD, enabling antemortem diagnosis of prion disease with high sensitivity (around 90%) and specificity (98–100%) in established cohorts.^5–14 Accordingly, expanded access to RT-QuIC testing in CSF has increased diagnostic confidence, supported earlier recognition of symptomatic patients, broadened appreciation of disease phenotypes, and improved prevalence estimates of CJD.^{6–10, 15} However, while RT-QuIC is indeed a good test, it is not perfect. False negative tests were reported in 48/497 (9.7%) of CSF samples from patients with autopsy-confirmed prion disease, including 31/439 (7.1%) patients with sCJD evaluated at the US National Prion Disease Pathology Surveillance Center.⁷ False negative rates may be considerably higher in patients with selected forms of genetic CJD, including mutations associated with Gerstmann-Straussler–Scheinker syndrome and fatal familial insomnia.¹⁶

It is unclear how well RT-QuIC test performance characteristics established in reference laboratories generalize to patients assessed in typical practice environments where diagnoses are established through longitudinal follow-up with or without autopsy confirmation. To address this question, we considered our experience within a cohort of consecutively encountered CJD patients assessed at multiple tertiary care centers in the United States. Clinical, laboratory, and pathological features were compared between CJD patients with positive and negative RT-QuIC tests. Findings were used to inform the diagnostic approach to RT-QuIC negative patients with suspected CJD.

Methods

Patients evaluated for CJD from 7/9/2013 to 10/7/2021 were identified through review of medical charts (Mayo Clinic Enterprise: Rochester, MN; Jacksonville, FL; and Scottsdale, AZ) or research records (Washington University in St. Louis: Saint Louis, MO) as previously reported.^{3, 10} Records were dual-reviewed to identify patients who met criteria for probable (neuropsychiatric disorder with positive CSF RT-QuIC; or rapidly progressive dementia with ≥1 of myoclonus, visual/cerebellar signs, pyramidal/extrapyramidal signs, or akinetic mutism; and consistent brain MRI) or definite CJD (pathology- or genetically-confirmed).¹⁷ One-hundred thirty-eight patients met defined criteria. Twenty-five patients did not undergo RT-QuIC testing and were excluded from further analyses. RT-QuIC testing for abnormal prions was performed on CSF from the remaining 113 patients at the National Prion Disease Pathology Surveillance Center (Cleveland, OH), as previously described.^{7, 9, 18} Protein 14–3-3 (qualitative measures by Western blot with an anti–14–3-3 beta polyclonal antibody; Abcam, Cambridge, MA) and total tau (quantitative measure by ELISA; Life Technologies, Carlsbad, CA) were performed in the same laboratory.⁷ For all patients, CSF was obtained by clinical teams in outpatient and inpatient settings utilizing standard clinical techniques. Procedures governing sample handling, processing, and shipment via clinical laboratories were not recorded in the clinical records.

Relevant clinical data were extracted and analyzed from included patients, including demographic information (age, sex, race), time from symptom onset to presentation, time from symptom onset to lumbar puncture, clinical features at presentation, diagnostic testing (MRI, EEG, CSF studies), and other relevant CSF studies (14–3-3, total tau). Death dates and results of neuropathological and genetic testing were recorded when available. A waiver of consent was granted by the Mayo Clinic Institutional Review Board to include deidentified data from Mayo Clinic patients. Patients at Washington University in St. Louis were consented and enrolled in prospective studies of rapidly progressive dementia. Study procedures conformed with the principles of the World Medical Association Declaration of Helsinki and were approved by the institutional review boards at the respective centers.

Differences between RT-QuIC positive and negative patients were analyzed using Fisher’s exact test for categorical variables and non-parametric Mann-Whitney U-tests for continuous variables. Statistical analyses were conducted using SPSS Statistics (IBM Corp., Version 25.0. Armonk, NY), with statistical significance established at p<0.05.

Results

Clinical symptoms, signs, and diagnostic tests were consistent with probable (n=64) or definite CJD (n=49) in 113 patients. No prions were identified via RT-QuIC assays in 13 patients (Table 1), corresponding to a sensitivity of 88.5% (100/113). RT-QuIC negative patients all had cognitive symptoms at presentation, with 11/13 (84.6%) patients presenting with rapid progressive dementia. Clinical features and the results of investigations are summarized in Table 2. RT-QuIC negative patients were younger at symptom onset (median 52.0 years, range 20.2–69.3) than RT-QuIC positive patients (median 66.1 years, range 21.9–82.8; p<0.001), with similar sex and racial distribution. Presenting symptoms and signs were similar between cohorts. Ataxia was less frequent in RT-QuIC negative (4/13, 30.8%) vs positive (61/100, 61.0%) patients, although differences did not reach predetermined levels of statistical significance (p=0.070). All RT-QuIC negative patients had brain MRI findings consistent with CJD. “Cortical ribboning” (increased T2/fluid-attenuated inversion recovery [FLAIR] and/or diffusion-weighted imaging [DWI] signal within the cortical ribbon) was more commonly observed in RT-QuIC negative (13/13) than positive patients (67/100; p=0.010). Although routine CSF studies were similar across cohorts, differences were observed in non-specific measures of neurodegeneration. Specifically, RT-QuIC negative patients demonstrated lower median total-tau levels (p=0.020) and lower rates of 14–3-3 positivity (p<0.001). Lower measures of neurodegeneration were not attributable to delays in CSF sampling. On the contrary, RT-QuIC negative patients presented later in the illness course (median 153 days following symptom onset, range 30–427) than RT-QuIC positive patients (median=47 days, range 0–882; p=0.003), and underwent CSF sampling even later (p=0.003).

Table 1:

Case details for RT-QuIC-negative patients with CJD.

Case/ Age, Sex	Days from Symptom Onset to,			Clinical Features and Investigations			CSF Studies						Autopsy Results
	Presentation	Lumbar Puncture	Death	Dominant Phenotype	MRI, Location of Abnormal Signal	EEG Findings	RBCs /mL	WBC /mL	protein mg/dL	total tau pg/ml	14–3-3 protein	RT-QuIC, 2nd LP	Molecular Subtype
A/63.8, M	92	122	1049	Cerebellar	Cortical	FED	0	0	55	982	−	+	NA
B/69.3, M	127	285	925	Cognitive	Cortical	NA	1	1	33	1154	−	NA	NA
C/45.0, F	30	189	269	Global	Cortical/ Subcortical	Slowing, PLED	0	1	32	699	−	NA	NA
D/61.7, M	335	615	Living	Cognitive	Cortical	Slowing, triphasic waves	91	1	36	5202	Equivocal	−	Living
E/67.8, F	335	335	Living	Cognitive	Cortical	Slowing	1881	1	59	3134	Equivocal	NA	Living
F/51.6, F	175	176	Living	Cognitive	Cortical/ Subcortical	Slowing	0	3	41	2120	+	+	Living
G/32.3, M	213	228	665	Motor	Cortical/ Subcortical	Slowing, FED	38	1	40	4001	+	NA	sCJD VV1
H/20.2, M	123	191	886	Psychiatric	Cortical/ Subcortical	Slowing, PLED	2	1	51	138	−	NA	sFI
I/49.1, F	183	185	305	Cognitive	Cortical/ Subcortical	Slowing	0	4	58	4001	+	NA	sCJD VV1
J/63.4, M	89	113	1548	Cognitive	Cortical/ Subcortical	Slowing	0	0	27	2517	Equivocal	NA	sCJD MM1
K/52.0, M	145	126	126	Cognitive	Cortical	Slowing	0	0	30	3048	Equivocal	NA	sCJD MV1
L/62.0, M	153	154	177	Global	Cortical/ Subcortical	Slowing	0	4	48	4681	+	NA	sCJD VV2
M/35.0, M	427	503	755	Cognitive	Cortical/ Subcortical	Slowing	0	1	22	754	−	NA	sCJD MM2

Dominant phenotype at presentations were assigned based on reported symptoms and signs at presentation as previously described.¹⁰ Cerebellar presentation: predominant ataxic gait and/or dysmetria. Cognitive presentation: predominant dysfunction in language, memory, executive, visual, or auditory function. Motor presentation: prominent pyramidal signs, extrapyramidal signs, corticobasal-like syndrome, or hyperkinetic movement disorder. Global presentation: near-simultaneous onset of symptoms across several domains (e.g., cognitive, cerebellar, motor).

+ = positive; − = negative

Abbreviations: M, male; F, female; CJD, Creutzfeldt-Jakob disease; FED, focal epileptiform discharges; NA, not available / not performed; PLED, periodic lateralized epileptiform discharges; sFI, sporadic fatal insomnia

Table 2:

Demographics, clinical features, and test results in RT-QuIC negative and positive patients with CJD.

Cohort Characteristics & Findings	Total Cases (n=113)	RT-QuIC Negative (n=13)	RT-QuIC Positive (n=100)	P-value
Patient Information
Age, years, median (range)	65.5 (20.2–82.8)	52.0 (20.2–69.3)	66.1 (21.9–82.8)	<0.001
Sex, male (%)	53 (46.9)	9 (69.2)	44 (44.0)	0.138
Race, white (%)	107 (94.7)	12 (92.3)	95 (95.0)	0.528
Days from symptom onset to presentation, median (range)	56.0 (0–882)	153.0 (30–427)	47 (0–882)	0.001
Disease duration, days, median (range)	166.5 (43–1732)	710.0 (177–1548) n=10	147.5 (43–1732) n=94	0.001
Days from symptom onset to CSF with RT-QuIC, median (range)	103.0 (8–1052)	189.0 (113–615)	84.0 (8–1052)	0.003
Clinical Features at Presentation
Cognitive, n (%)	103 (91.2)	13 (100)	90 (90)	0.602
Behavioral, n (%)	43 (38.1)	6 (46.1)	37 (37)	0.554
Sensorimotor, n (%)	80 (70.8)	8 (61.5)	72 (72)	>0.99
Ataxia, n (%)	65 (57.5)	4 (30.8)	61 (61)	0.070
Visual, n (%)	26 (23.0)	2 (15.4)	24 (24)	0.729
Constitutional, n (%)	16 (14.2)	2 (15.4)	14 (14)	>0.99
Rapidly progressive dementia, n (%)	96 (85.0)	11 (84.6)	85 (85)	>0.99
CSF Studies
RBC, median (range)	1.0 (0–4000)	0 (0–1881)	1.0 (0–4000), n=96	0.358
WBC cells/mcL, median (range)	1.0 (0–41)	1.0 (0–4)	1 (0–41), n=95	0.929
Protein mg/dL, median (range)	42.0 (20–151)	40.0 (22–59)	42.0 (20–151), n=96	0.328
Glucose mg/dL, median (range)	64.0 (46–141)	62.0 (56–83)	64.5 (46–141), n=94	0.437
CSF tau >1150 pg/mL, n (%)	100/111 (90.1)	9 (69.2)	91/98 (92.9)	0.077
Tau pg/mL, median (range)	4001 (27–23586)	2517 (138–5202)	4001 (27–23586), n=98	0.020
14–3-3 positive, n (%)	81/107 (75.7)	4 (30.8)‡	77/94 (77)	<0.001
Diagnostic Studies
EEG, n (%)	102 (90.3)	12 (92.3)	90 (90.0)	>0.99
Normal EEG, n (%)	12/102 (11.8)	1/12 (8.3)	11/90 (12.2)	>0.99
Slowing, n (%)	83/102 (81.4)	11/12 (91.7)	72/90 (80.0)	0.456
Epileptiform discharges, n (%)	7/102 (6.9)	2/12 (16.7)	5/90 (5.6)	0.191
Periodic discharges/PLEDS, n (%)	23/102 (22.5)	2/12 (16.7)	21/90 (23.3)	>0.99
MRI consistent with CJD, n (%)†	108 (95.6)	13 (100)	95 (95.0)	>0.99
Cortical ribboning >2 areas	80 (70.8)	13 (100)	67 (67.0)	0.010
Involvement of deep nuclei	58 (51.3)	7 (53.8)	51 (51.0)	>0.99

PLEDS, periodic lateralized epileptiform discharges; RT-QuIC, real-time quaking induced conversion assay for prion disease

^†High signal in caudate/putamen on magnetic resonance imaging (MRI) or at least two cortical regions (temporal, parietal, occipital) either on diffusion-weighted imaging (DWI) or fluid attenuated inversion recovery (FLAIR) imaging.

^‡14–3-3 was equivocal in 3 RT-QuIC negative patients

RT-QuIC testing was repeated in 3 RT-QuIC negative patients, returning positive in 2/3 cases. One patient was a 63-year-old male (Case A) who presented 92 days after symptom onset and received his first RT-QuIC test 30 days later (day 122). Repeat RT-QuIC on day 154 was positive. The second patient was a 51-year-old female (Case F) who presented 175 days from symptom onset and had initial RT-QuIC testing performed the following day. Repeat RT-QuIC on day 237 was positive. The patient with persistently negative RT-QuIC was a 61-year-old male (Case D) who underwent an initial lumbar puncture 553 days following symptom onset. His initial CSF was contaminated with 91 red blood cells. Diagnostic lumbar puncture was repeated 48 days later (day 663) and was again contaminated with red blood cells (28 mg/dL). CSF total-tau levels increased from 5202 to 7672 pg/mL; 14–3-3 was positive on the second lumbar puncture. Autopsy and genetic testing were not available for any of these patients. No patient had a family history of prion disease.

At the time of analysis, 104 patients had died (92.0%), including 10/13 (76.9%) RT-QuIC negative patients. Seven RT-QuIC negative patients underwent autopsy. One patient had a diagnosis of sporadic fatal insomnia. The remaining six patients had sCJD with MM1, MM2, VV1 (two patients), and MV1 molecular subtypes (Supplemental Table). Patients with the VV1 subtype were over-represented in the RT-QuIC negative group (p=0.056). Median survival (symptom onset to death) was substantially longer in RT-QuIC negative (710 days, range 177–1548) vs positive patients (147.5 days, range 43–1732; p=0.001).

Discussion

RT-QuIC is a sensitive measure of prions in CSF. Diagnostic performance in our clinical cohort paralleled that reported in national surveillance centers and research cohorts using second-generation RT-QuIC assays.^5–14 Yet, as our experience shows, false negative results do occur in practice. It is important for clinicians to recognize scenarios in which RT-QuIC may underperform and to leverage other diagnostic tests and tools to make the correct diagnosis, support accurate counselling, and optimize care for the patient with unexpectedly negative RT-QuIC results.

In our series, false negative RT-QuIC tests were more frequent in younger patients and those with lower concentrations of CSF total-tau and 14–3-3. Similar findings were reported in a large autopsy-confirmed series conducted at a national surveillance center for prion disease, which included 48 RT-QuIC negative (9.7%) and 449 positive patients with definite prion disease. In this series, younger age was independently associated with false negative testing in multivariate analyses. Frequency of 14–3-3 positivity and total tau levels were also lower in RT-QuIC negative patients, with greater delays to first presentation and diagnostic evaluation associated with RT-QuIC negative status in univariate comparisons.⁷ Collectively, these observations suggest the symptomatic course of prion disease may be more indolent in RT-QuIC negative patients, recognizing that CSF measures of 14–3-3 and total tau are less specific biomarkers of neurodegeneration,¹⁹ and that time to presentation may be inversely correlated with the speed of prion propagation.

Disease-specific factors also influence assay performance. Lower RT-QuIC sensitivity is reported in patients with VV1 and MM2 molecular subtypes of sCJD,⁷ selected genetic forms of CJD (e.g., Gerstmann-Straussler-Scheinker syndrome),^{16, 20} and patients with fatal familial insomnia (sporadic and inherited subtypes).^{7, 20} These patients also have a younger age at onset (VV1 and MM2 or sporadic fatal insomnia), a longer symptomatic course, lower median total tau levels, reduced sensitivity of the 14–3-3 assay, and are less likely to present with ataxia.^{7, 9} These findings explain at least 4/13 (30.8%) of the RT-QuIC negative patients encountered in our cohort. It is unknown whether these forms of prion disease associate with lower levels of abnormal prion proteins, conformational shapes that are less prone to amplification and detection via RT-QuIC, or some other change that compromises assay performance. Nevertheless, when an unexpected RT-QuIC result is encountered, the astute physician must consider atypical subtypes.

Finally, preanalytic factors may alter assay performance, as evidenced by the trend towards false negative tests reported in non-colorless (i.e., bloody) CSF samples tested via the US National Prion Disease Pathology Surveillance Center.⁷ Similar findings were noted in another study, which also questioned whether elevated CSF protein (>45 mg/dL) and albumin (>35 mg/dL) might interfere with RT-QuIC seeding or thioflavin florescence (critical to assay performance²¹).¹³ Assessing clinicians should bear these factors in mind, taking care to send colorless samples for RT-QuIC testing whenever possible.

These findings inform the approach to the evaluation of patients with suspected CJD with unexpectedly negative RT-QuIC. An unexpectedly negative test should prompt careful consideration (or reconsideration) of the differential diagnosis, weighing available clinical and paraclinical information. In our cohort, all RT-QuIC negative patients had MR neuroimaging findings that were highly consistent with CJD, including increased signal on T2/FLAIR and/or DWI sequences in multiple cortical, striatal or subcortical structures.²² This observation emphasizes the value of this accessible and widely available diagnostic test—particularly when reviewed by neurologists and neuroradiologists experienced in the diagnosis of CJD.²³ Yet despite the high sensitivity of brain MRI for the diagnosis of CJD, features compelling for CJD continue to be under-recognized, especially in patients in whom signal abnormality is present in limited brain areas.²⁴ For this reason, MRI results should be interpreted together with findings from CSF tests, including measures of total tau and 14–3-3. Indeed, although median total tau levels were lower in RT-QuIC negative patients in this cohort, levels exceeded the threshold commonly associated with CJD (>1150 pg/mL)^{7, 19, 25, 26} in 9/13 RT-QuIC negative cases. The same cannot be said for 14–3-3, which was positive in only 4 RT-QuIC negative CJD patients and equivocal in three other patients. An unexpectedly negative test should prompt careful consideration (or reconsideration) of the differential diagnosis, weighing available clinical and paraclinical information. Indeed, repeat RT-QuIC testing was positive in two patients with probable CJD in this series who were retested 32 (Patient B) and 61 days (Patient F) after an initial negative test. Genetic counselling and targeted assessment for variants associated with familial prion disease (and other inherited causes of dementia) should be recommended in a patient-specific manner,³⁰ particularly when a family history of atypical dementia or early death is suspected.

Our results are subject to several limitations. Patients with probable CJD were identified in accordance with updated diagnostic criteria. In patients with negative RT-QuIC, these criteria require supportive findings on CJD on MRI, EEG, or CSF (e.g., positive 14–3-3) consistent with CJD.¹⁷ It is therefore likely that RT-QuIC negative CJD cases were underestimated, recognizing the higher burden of proof required to establish a diagnosis of probable CJD in these cases. Additionally, although we included all known cases of CJD assessed at our centers, the relatively small number of false negative cases limited our ability to detect more modest differences in RT-QuIC negative and positive cases. The even smaller number of samples submitted for retesting leaves open the possibility that repeat testing may have yielded “true positive” results in these selected cases. While the factors found to be associated with false negative RT-QuIC testing are not novel, our findings replicate those from national surveillance centers, and research settings,^{7–9, 13, 14} suggesting that estimates of RT-QuIC assay sensitivity may be generalized to clinical settings. This is particularly notable recognizing that CSF collection, processing, and handling may be more variable in clinical practice and submitted samples more likely to include blood or other excipients that could alter assay performance. Larger series pooling data from greater number of centers and controlling for important preanalytic variables (e.g., handling of CSF) may further inform the features that associate with false negative RT-QuIC testing and the optimal approach to these patients.

Conclusion

The sensitivity of RT-QuIC in patients with CJD approaches 90% in clinical practice, reinforcing that RT-QuIC is a good test for the diagnosis of CJD. An unexpectedly negative RT-QuIC test requires the assessing clinician to carefully (re)consider the differential diagnosis, while leveraging additional diagnostic tests to make the correct diagnosis and optimize care for the patient with suspected CJD.

Material Availability

Anonymized study data will be shared pending review of a request from qualified individuals by the corresponding author.