Journal of Neurology
ORIGINAL COMMUNICATION
Application of real‑time quaking‑induced conversion in Creutzfeldt–Jakob disease surveillance
Peter Hermann1 · Matthias Schmitz1,2 · Maria Cramm1 · Stefan Goebel1 · Timothy Bunck1 · Julia Schütte‑Schmidt1 · Walter Schulz‑Schaefer3 · Christine Stadelmann4 · Jakob Matschke5 · Markus Glatzel5 · Inga Zerr1,2
Received: 25 November 2022 / Revised: 22 December 2022 / Accepted: 23 December 2022
© The Author(s) 2023
Abstract
Background Evaluation of the application of CSF real-time quaking-induced conversion in Creutzfeldt–Jakob disease surveillance to investigate test accuracy, influencing factors, and associations with disease incidence.
Methods In a prospective surveillance study, CSF real-time quaking-induced conversion was performed in patients with clinical suspicion of prion disease (2014–2022). Clinically or histochemically characterized patients with sporadic Creutzfeldt– Jakob disease (n=888) and patients with fnal diagnosis of non-prion disease (n=371) were included for accuracy and association studies.
Results The overall test sensitivity for sporadic Creutzfeldt–Jakob disease was 90% and the specificity 99%. Lower sensitivity was associated with early disease stage (p=0.029) and longer survival (p<0.001). The frequency of false positives was significantly higher in patients with inflammatory CNS diseases (3.7%) than in other diagnoses (0.4%, p=0.027). The incidence increased from 1.7 per million person-years (2006–2017) to 2.0 after the test was added to diagnostic the criteria (2018–2021).
Conclusion We validated high diagnostic accuracy of CSF real-time quaking-induced conversion but identified inflammatory brain disease as a potential source of (rare) false-positive results, indicating thorough consideration of this condition in the differential diagnosis of Creutzfeldt–Jakob disease. The surveillance improved after amendment of the diagnostic criteria, whereas the incidence showed no suggestive alterations during the COVID-19 pandemic.
snip...
Development of CJD incidence in Germany (2006– 2021)
The cumulative incidence of sCJD has increased from 1.7 per million person-years 2006–2017 to 2.0 per million person-years in 2018–2021, when diagnostic criteria including RT-QuIC were applied prospectively (Fig. 3). These data also include sCJD cases that were classified without RTQuIC analyses (based on 14-3-3, MRI, and EEG only) as well as autopsy results from cases without available clinical data. In some RT-QuIC-positive cases, no further or no sufficient clinical information was available to the CJD Surveillance group. These cases were indicated as “unclarified” throughout this manuscript. Including these patients resulted in a cumulative incidence of 2.1 per million person-years (2018–2021). We could not observe any suggestive alteration of sCJD incidence in the years of the COVID-19 pandemic 2020 (2.10) and 2021 (2.01) compared to the preceding year 2019 (2.08), the first year in which all cases were systematically classified according to the amended criteria.
snip...
Discussion
CSF RT-QuIC has become the gold standard in the laboratory-based diagnosis of sCJD [29]. It is currently applied as a solitary criterion within the biomarker-set of diagnostic protocols [12] and as a “confirmatory” test after CJD typical results from other biomarker analyses [30]. Here, we present comprehensive data from a well-established surveillance system including all patients with diagnosis of sCJD (n=888) and a highly specific cohort of CJD mimics (n=371).
We observed a sensitivity of 90% for sporadic sCJD, which is comparable to some reports from surveillance centers using 1st generation RT-QuIC [31, 32] and slightly below the sensitivity of a modified protocol called 2nd generation RT-QuIC (IQ), which was introduced in 2015 [33] and showed a sensitivity ranging from 92 to 96% [33–35]. Several factors that may influence the test sensitivity have been proposed in the past. We observed no significant differences between true and false-positive sCJD patients regarding age and sex, in contrast to a previous study that identified false-negative results to be associated with lower age and female sex [19]. We can only speculate on the reasons for this discrepancy, but it could possibly be associated with different test protocols or with the investigated sCJD cohorts. In that study, only autopsy confirmed cases were evaluated and we observed differences of the age and the sex distribution between probable and definite cases in our cohort (Table 2). We observed no difference of the overall test sensitivity between probable and definite patients in our cohort, though. However, we could replicate an association between longer disease duration and test negativity, and showed that false negative RT-QuIC is associated with early disease stage.
We also validated high sensitivity for the most frequent MM/MV1, VV2, and MV2 sCJD subtypes [19, 34, 35], whereas sensitivity seems to be lower in the rare MM2 and VV1 subtypes [19, 34]. In some studies [34], sensitivity was slightly higher than in ours, possibly due to different substrates or test protocols. On the other hand, investigated case numbers of these subtypes were rather low in all studies (around 10 or less in each), which may not allow to draw final conclusions. An explanation for the low sensitivity may be that MM2C and VV1 subtypes show predominant cortical PrPSc pathology in early disease stages [6] and slower disease progression than most other sCJD subtypes, possibly resulting in less amount of PrPSc in the CSF. This would be in line with evaluations in genetic prion diseases that reported low sensitivity in entities with slow disease progression (GSS) or pathology restricted to defined structures (brainstem and thalamus in FFI, cerebellum in GSS) in early disease stages [36].
Regarding the specificity of RT-QuIC, previous analyses showed an excellent accuracy of CSF RT-QuIC of about 99% or higher [12]. Our data from eight years of clinical application indicated a specificity of 99% but for the first time, we evaluated test-related and clinical data in a series of five false-positive cases. Four of them were diagnosed with immune-mediated encephalitis and the rate of false positives was significantly higher than in other diagnostic groups (p=0.027). In addition, single (one of the three) false-positive signal increases were also significantly more frequent among differential diagnoses with inflammatory pathophysiologic background (p=0.001), suggesting a potential causal relationship between encephalitis and false-positive RT-QuIC results. Of course, analytical and pre-analytical factors cannot be excluded. In the literature, only 10 cases of false-positive RT-QuIC or Endpoint-QuIC with clear diagnosis have been reported (see Table 5). They were diagnosed with vascular dementia [32], Alzheimer’s disease [25], mixed dementia [19], and tauopathies [35, 37], immune-mediated encephalitis [20, 35, 38], and amyloid associated vasculitis [39].
The possibility of a higher likelihood of false-positive RT-QuIC among encephalitis patients is an important issue because many inflammatory encephalopathies are highly treatable and may represent the most important clinical mimics of CJD and causes of rapidly progressive dementia [40]. On the other hand, all false-positive patients showed either clinical or CSF characteristics that pointed to other diagnoses than CJD, indicating that consideration of factors such as inflammatory signs in the CSF may improve the specificity of an RT-QuIC-based clinical diagnosis. Total CSF Tau protein may also give additional clues because of better specificity in the discrimination of CJD and acute encephalopathies than 14-3-3, but Tau was not available for the false-positive patients. On the other hand, total-tau may also be extremely elevated in encephalitis [41] due to ongoing severe neuronal damage. More important, none of the patients showed CJD-typical MRI. In our autopsy series, no patient received incorrect ante-mortem diagnosis of CJD based on RT-QuIC positivity. We identified only five false-positive results in eight years of RT-QuIC application in a sum of 4599 patients. However, clinical information was only available for 371 control patients, leading to the reported specificity of 99%.
Further investigations have to validate our findings about false-positive RT-QuIC and investigate potential mechanisms. So far, previous studies have not found association of RT-QuIC efficiency and neuronal damage markers such as total-tau and proteins 14-3-3 in sCJD patients [42]. Total PrP was also not associated with seeding efficiency in sCJD [42] but has not been investigated as a factor for false-positive RT-QuIC, yet. On the other hand, total PrP was not shown to be significantly altered in encephalitis compared to cerebral ischemia or control patients [43]. Another potential reason may be the influence of factors in the CSF that are directly linked to neuro-inflammation. Epileptic activity in encephalitis patients was also discussed as a cause for false-positive results [21]. Our data did not allow the evaluation of the presence of seizures in relation to lumbar puncture in control patients, but patients with primary diagnosis of seizures or status epilepticus caused by idiopathic epilepsy syndromes, or reversible conditions such as alcohol withdrawal showed a low frequency of positive test replicates (one in 30 patients). However, the mechanisms for false-positive results may be related to the CSF. RT-QuIC from other body tissues such as olfactory mucosa [44, 45] are an alternative clinical test and should be investigated in future studies on the specificity of RT-QuIC in neuro-inflammatory diseases. Clarification of the reasons for false-positive PrPSc RT-QuIC reactions may also be highly relevant for the application of other protein amplification assays such as α-Synuclein RT-QuIC. So far, only very few false-positive results have been reported and were associated with Wernicke’s encephalopathy, Alzheimer’s disease, and encephalitis [46, 47].
Although our study provides comprehensive data on clinical experience with RT-QuIC, the study has naturally some limitations. First, our test protocol [24] uses chimeric hamster-sheep recombinant PrP as substrate, whereas most other centers are using hamster recombinant PrP [32, 33]. Some centers showed that IQ-CSF RT-QuIC may have a superior sensitivity for sCJD compared to previous protocols [30, 45, 48]. It remains unclear, whether IQ RT-QuIC underlays the same or similar confounders for PrP seeding as our protocol but international ring trials have shown that RT-QuIC results are highly concordant among different test centers and test protocols [25, 49]. Our protocol has been well established over years and we have achieved a high level of experience to perform the test in a reliable and reproducible way.
Regarding the sensitivity, further histochemical characterization of sCJD was only available in a limited number of cases (n=161), discouraging reliable conclusions on the sensitivity of rare (MM2, VV1) sCJD subtypes. Similarly, Codon 129 was analyzed in a rather small subset of patients (n=114). Lastly, our surveillance data includes a number of uncharacterized cases with suspected prion disease or positive CSF RT-QuIC. In these cases, further clinical data were not available and thus, we cannot exclude additional false-positive or false-negative test results in this group. Our cohort may be prone to an according selection bias. On the other hand, we assume that this bias might rather be less relevant than in retrospective case–control studies or in studies including only autopsy-confirmed cases.
As a secondary outcome, we observed an increased overall sCJD incidence in Germany after inclusion of the test in the clinical diagnostic protocol. In this context, we assume that the increase from 1.7 (2009–2017) to 2.0 (2018–2021) person-years is a result of the improved clinical detection of early sCJD cases (based on positive RT-QuIC) as previously suggested or observed by our group [18] and others [13, 17, 19]. This effect was also described in the context of previous criteria modifications [50]. Another interesting observation was the apparent lack of an alteration of the overall sCJD incidence during the Covid-19 pandemic. Annual numbers of positive RT-QuIC results remained stable between 2019 and 2021 (Fig. 3). Although we cannot exclude influence of viral infections on genesis or course of prion diseases, our surveillance data do not suggest an immediate causal relationship between COVID-19-related factors and CJD on the population level.
Conclusion
Chimeric PrP CSF RT-QuIC is an accurate diagnostic tool for the differential diagnosis of sCJD. If the test is interpreted in the context of a complete diagnostic work-up, it may provide an extremely high level of ante-mortem diagnostic certainty. RT-QuIC negativity combined with absence of CJD-typical results in 14-3-3 analysis and MRI indicates extremely low likelihood of sCJD. The routine application of RT-QuIC improves CJD surveillance and leads to a formal increase of the disease incidence. However, the sensitivity is influenced by disease stage and disease subtype. False positive results may occur and clinicians have to be aware of this possibility. In cases with ambiguous clinical presentation, we recommend consideration of other diagnostics, in particular MRI, and repetitive RT-QuIC analyses from the same sample to exclude the influence of analytical factors. Nonetheless, a consecutive lumbar puncture, at best after therapeutic intervention, may be necessary to detect false positivity based on pre-analytical or disease-related factors. In this context, RT-QuIC from other body tissues such as olfactory mucosa may be an alternative, if available. Inflammatory CNS disease, especially immune-mediated encephalitides, should always be considered as potential clinical and laboratory mimics of CJD. Although general comparability of different RT-QuIC protocols and substrates have been shown, our pilot findings need to be verified through studies using other body tissues and test protocols such as IQ.
Supplementary Information The online version contains supplementary material available at doi.org/10.1007/s00415-022-11549-2 .
Acknowledgements We would like to thank all collaborating health care institutions, hospitals, physicians, patients, and families that provided information for diagnostic classifications. In addition, we acknowledge the Center for Neuropathology and Prion Research at the Ludwig-Maximilian-University in Munich, Germany for performing neuropathological investigations in several cases of suspected prion disease.
Funding Open Access funding enabled and organized by Projekt DEAL. This study was funded by the Robert-Koch-Institute through funds from the Federal Ministry of Health; grant number 1369-341. Data availability All manuscript-related data is available and will be provided upon reasonable request.
Declarations
Conflicts of interest The authors declare that they have no relevant financial or non-financial interests to disclose.
Keywords Creutzfeldt–Jakob disease · Prion · Diagnosis · RT-QuIC · Epidemiology