Abstract
Background: In cell culture experiments, flupirtine maleate (FLU), a triaminopyridine compound, was able to protect neuronal cells from apoptotic cell death induced by prion protein fragments and β-amyloid peptides. As FLU is a clinically safe drug, the authors started a double-blind placebo-controlled study in patients with Creutzfeldt–Jakob disease (CJD).
Methods: Twenty-eight patients with CJD were randomized to an oral treatment with either FLU (n = 13) or matching placebo (PLA; n = 15). For inclusion and continuing the study, the patients had to achieve at least 50% in two of the subscales of the dementia tests employed. A battery of standardized questionnaires was employed to monitor the progression of the disease. The main outcome variable was the cognitive part of the Alzheimer’s Disease Assessment Scale (ADAS-Cog); the difference between baseline and the best score under treatment was defined as the primary efficacy variable for hypothesis testing.
Results: CJD types were homogeneously distributed among the treatment groups. Patients treated with FLU showed significantly less deterioration in the dementia tests than patients treated with PLA. The mean change in ADAS-Cog (baseline to best) was +8.4 (±15.3) in the FLU group and +20.6 (±15.1) in the PLA group (p = 0.02, one-sided t-test).
Conclusions: FLU has beneficial effects on cognitive function in patients with CJD. These positive results also may suggest a treatment potential of FLU in other neurodegenerative disorders. However, further studies are necessary.
Pharmacologic interventions for patients with Creutzfeldt–Jakob disease (CJD) are especially hampered by the lack of early clinical diagnosis and objective follow-up parameters for monitoring disease progression and therapeutic benefit. Especially in humans, survival time might be an ambiguously interpretable variable, as quality of life and consciousness of a patient might be so badly impaired under therapy that a treatment to prolong life might no longer be justified.
Currently, diagnosis of CJD is rendered according to clinical symptoms and can be supported by elevated levels of so-called surrogate markers in the CSF or serum.1,2⇓ The diagnosis of variant CJD can be rendered by demonstrating the pathologic isoform of prion protein (PrP) in tonsil biopsies.3 However, these tests are usually conducted when the patient is already in a symptomatic phase. This means that currently pharmacotherapy should be administered only to prevent formation of aggregated PrP or suppress damage by these aggregates.
A variety of compounds have been tested in cellular and animal models.4–6⇓⇓ Flupirtine maleate (FLU) is a well-tolerated, centrally acting, nonopioid analgesic, which has been found to display cytoprotective activity in vitro and in vivo on neurons induced to undergo apoptosis.7,8⇓ Neurotoxicity caused by the PrP106-126 fragment was greatly reduced following coincubation with FLU.9
Owing to the small number of patients and the varying dynamics of CJD,10 only a double-blind study will elucidate the role of any drug. As FLU is a clinically well-established drug, we performed such a study on 28 patients with CJD. The main outcome parameter was the cognitive part of the Alzheimer Disease Assessment Scale (ADAS-Cog) score.
Patients and methods.
From September 1997 to September 2001, patients were recruited from among those reported to the German National CJD Surveillance Unit in Goettingen. Initially, every patient was visited by a research physician and examined using a standardized protocol. According to the clinical criteria, suspected cases of CJD were classified as “probable,” “possible,” or “other” cases.11
Patients diagnosed as “probable” and fulfilling the inclusion criteria were referred to the research physician in charge of this study. After informed consent, patients were admitted to this study always in co-operation with local doctors in charge (figure 1).
Figure 1. Flow chart of Creutzfeldt–Jakob disease trial.
The study was approved by the local ethics committee in Goettingen.
CSF and MRI findings.
CSF analysis for 14-3-3 protein, tau protein, and S-100B was performed according to standard procedures.1 Furthermore, all MRI were re-evaluated for lesions typical of CJD.
Mental state testing and inclusion and exclusion criteria.
At regular intervals of 2, 4, 8, 12, 16, and 20 weeks after treatment start, patients were examined using a battery of standardized mental state tests. For inclusion and continuation in the study, patients had to achieve a score of at least 50% in 2 of 12 subtests of the dementia tests (ADAS-Cog, Goettingen CJD Dementia Test [GoeCJDDT]).
The mental state tests comprised the ADAS, the Mini-Mental State Examination (MMSE), and the GoeCJDDT. The test protocol always started with the ADAS. If patients were too exhausted after this test, MMSE and GoeCJDDT were no longer performed.
The main outcome variable was the cognitive part of the ADAS (ADAS-Cog). In this test, the maximum score of 70 points indicates the poorest performance.12 The MMSE was performed according to the published protocol,13 with a score of 30 points signifying the best performance.
The GoeCJDDT was newly developed to take into account the particular difficulties of testing CJD patients. These patients often have ataxia, visual impairment, and a complex array of movement disorders, which limits their ability to perform the usual cognitive test. The GoeCJDDT comprises tasks testing long-term and short-term memory, aphasia, attention span, and executive functions. A total score of 33 points indicates the best performance.
Additionally, a mental state questionnaire was completed with the assistance of relatives of the patients for an “outside” rating.14 This rating consists of two parts: one focusing on cognitive function and the other on noncognitive performance. A total of 110 points indicates the best performance for the cognitive subscales, and a maximum of 175 points can be attained for the noncognitive subscales.14
Medication.
Patients were randomized to a double-blind oral treatment with either FLU (n = 13) or matching placebo (PLA; n = 15). After all eligibility criteria were fulfilled, treatment commenced generally on the day of the mental state testing. Starting with 1 capsule/day (each containing 100 mg of FLU or matching PLA), the maintenance dosage was 3 to 4 capsules/day. This dose was reached within 3 days. Medication was stopped when patients no longer fulfilled the inclusion criteria. Adverse events and any concurrent medication were recorded at each visit.
Neuropathology.
Twenty-four of 28 patients died up to closure of the database. CJD was neuropathologically verified according to standard methods.15 Additionally, PrP typing was done. A genetic examination of the coding region of the PrP gene was performed to screen for known and unknown mutations as well as to determine the polymorphism at codon 129.10 In two patients, a mutation was detected in the PrP gene. These two cases clinically presented as sporadic CJD. One patient was in the FLU and the other in the PLA group.
Statistical analysis.
The confirmatory comparison of groups was based on the change from baseline to the best result in ADAS-Cog under treatment. If assessments could no longer be performed owing to progression of disease, imputation using the worst possible score was performed to reflect the extremely poor cognitive capacity of the patients.
A difference of 4 score points in the ADAS-Cog between FLU and PLA was considered to be clinically relevant.16 Assuming SD = 7, α = 0.05, and a power of 70%, a one-sided t-test led to a sample size estimate of 30 patients/group.
A planned interim analysis was performed under blind conditions after treatment of 21 patients based on α adjustment according to O’Brien and Fleming.17 An independent study review committee recommended continuation of the study. At the planning stage of the study, it was expected to include about 50% of patients with dementia of other origin. As diagnostic methods improved during the course of the trial and according to the study review committee, the sample size of about 30 patients in total was regarded as sufficient.
The primary variable was compared for confirmation using a one-sided t-test at an α level of 0.048 (adjusted because of the interim analysis performed per study protocol) based on the intention-to-treat (ITT) analysis set. For sensitivity purposes, the primary variable was analyzed also for the per-protocol (PP) cohort and without applying any imputation strategy. Life table methods (Kaplan-Meier curves and log-rank tests) were applied on the duration of treatment and the time from start of treatment to death. The PP cohort was defined as a subset of the ITT population, including only subjects who were treated according to protocol, that is, showing no major protocol violation(s).
Results.
Thirteen men and 15 women with CJD were included in the study (table 1). Patients were between 35 and 74 years. One patient (PLA) discontinued the study without any efficacy data after 3 days owing to lack of tolerability and withdrawal of consent. This patient was included in the analysis of safety. The individual blinding was broken for one patient from each group, and they were consequently excluded from the PP cohort. One patient had urticaria and the second gastrointestinal bleeding. Additionally, for one patient of the FLU group, a muscle weakness was reported that did not lead to a change of study medication. Inspection of listings revealed no doubt about compliance.
Table 1 Patients’ characteristics at baseline
Diagnostic data.
All patients fulfilled clinically the diagnostic criteria for sporadic CJD. One patient showed initially a negative 14-3-3 western blot, but a control was positive; tau protein and S-100B protein were also above the cut-off for the diagnosis of CJD, and the MRI was typical for CJD. This case was later neuropathologically verified as sporadic CJD.
S-100B and tau protein were increased in CSF of all patients (see table 1). In three patients, tau protein was below the cut-off of 1,300 pg/mL but still over 1,000 pg/mL, and CJD was later neuropathologically verified.
Genetic investigations, polymorphism, and neuropathology.
Although all cases presented as sporadic CJD, two patients were later verified as genetic CJD cases. One patient had an E200K mutation and was included in the PLA group. The other patient, who had a D178N mutation with a methionine–valine (MV) polymorphism at codon 129 and the D178V mutation in coupling with valine at 129, was included in the FLU group.
In the PLA group, two patients were homozygous for methionine at codon 129 (MM), six patients were homozygous for valine (VV), and seven patients were heterozygous at codon 129 (MV). The FLU group comprised two patients with MM, four with VV, and seven with MV.
Twelve patients were neuropathologically verified as having sporadic CJD. PrP-typing data were available for eight patients.10 One patient was MV type 1, four patients were MV type 2, and three patients were VV type 2.
In the FLU group, one patient had one APOE allele ε4 (APOE-4), and in the PLA group, four patients had an APOE-4 allele.
Neuropsychological follow-up.
The median treatment time for the FLU group was 29 days for the ITT cohort (range 11 to 85 days, mean 44 days). The median for the PLA group was 20.5 days (range 8 to 86 days, mean 37 days). The comparison by a two-sided log-rank test on the duration of treatment resulted in a p value of 0.63.
ADAS-Cog.
The FLU group performed statistically significantly better with regard to the primary efficacy variable (table 2). Best follow-up results were usually obtained at the second or third visit. No postbaseline ADAS-Cog scores were obtained in five patients (one in the FLU group and four in the placebo group); hence, imputation was needed to reflect progressive disease.
Table 2 Change in neuropsychological test battery
For the ITT analysis, the mean ADAS-Cog score at baseline was 28.2 points for the FLU group. Patients in the PLA group performed slightly better at the first visit, with a mean ADAS-Cog score of 25.1 points. Patients under FLU increased in their mean ADAS-Cog score by 8.4 points, whereas patients under PLA increased by 20.6 points in mean.
In the ADAS-Cog subtests, patients under FLU performed significantly better in the memory and orientation items, whereas speaking items were not significantly different compared with the increase in the PLA group.
MMSE and GoeCJDDT.
Patients in the FLU group started with a mean MMSE score of 19.2 points (see table 1). Patients in the PLA group had a test score of 20.5 points. The decrease of the MMSE score was 3.3 points in the FLU group and 8.0 points in the PLA group. This just failed the level of significance (p = 0.07). A similar result was obtained for the GoeCJDDT (p = 0.06; see table 2).
Kessler Dementia Scale score.
In the cognitive part of this relative rating test (Kessler–Cognitive), the decrease was significantly more pronounced in the PLA group than in the FLU group. The decrease in the score was 13.8 in the FLU group. In the PLA group, the decrease was 34.6 points.
For the noncognitive part, decreases of 16.8 (SD 35.0) were observed in the FLU group and of 46.9 (SD 48.6) in the PLA group (p = 0.04; see table 2).
Survival analysis.
The median survival time in the FLU group was 107 days (mean 141 days) vs 106 days (mean 97 days) in the PLA group (figure 2). No relevant difference in the time from start of treatment to death was detected (p = 0.19, log-rank test, two tailed). The median time from discontinuation of the drug to death was 69 days (mean 97 days) in the FLU group and 57.5 days (mean 63 days) in the PLA group (p = 0.26).
Figure 2. Survival analysis (time to death of intention-to-treat cohort). PLA = placebo; FLU = flupirtine.
Discussion.
Although the ADAS-Cog was not designed for investigating CJD patients, it was found that patients could be easily followed up using this questionnaire. In this study, it was found that FLU has a significant effect on the cognitive function of patients with CJD. The supposed difference in the changes in ADAS-Cog of 4 points to reach a clinically relevant level was surpassed. A tendency toward better results in the FLU group was seen in the MMSE and GoeCJDDT, too. Apart from the possibility that these tests were not influenced by the medication, other reasons might be that patients were more exhausted when these tests were performed or that the total score of these tests was too low to demonstrate a more pronounced effect. The caregiver’s impression was also significantly better in the FLU group. The study was designed in such a manner that survival was not the primary endpoint, as this might be an ambiguous variable for patients with dementia and especially CJD. Nonetheless, it is important to state that the survival time analysis did not show a relevant difference in this study.
Interestingly, patients with MV and VV at codon 129 and type II PrPSC were overrepresented in our study, as patients with sporadic CJD mainly present with MM at codon 129. The course of disease of patients with MM at codon 129 is more rapid as compared with patients with other polymorphisms,10 shortening the interval between diagnosis and possible treatment options. It is clear that the main challenge in CJD is early diagnosis, which otherwise limits therapeutic options. It remains open whether a combination of treatment strategies and compounds will result in a better outcome for the patient. At the moment, patients with genetic presymptomatic CJD and patients with a high risk of developing CJD (e.g., iatrogenic cases) might be especially receptive to therapy with FLU. It has to be said that FLU does not cure CJD patients, but it is hoped that this first study may lead to more effective therapies. In further double-blind studies, FLU might now be used instead of a placebo, although our results certainly do not preclude the use of placebos in future trials.
It is proposed that FLU acts neuroprotectively by up-regulation of the proto-oncogene bcl-2 and a normalization of the glutathione level.9 A neuroprotective effect of FLU has already been described in other cell culture and animal models.7,18⇓ However, further research in more advanced cell culture models and animal models is necessary to elucidate the mode of action of flupirtine in transmissible spongiform encephalopathy. Interestingly, a similar neuroprotective effect was seen when FLU was coincubated with β-amyloid peptides.8 This result suggests a treatment potential of FLU in Alzheimer disease.
Acknowledgments
Supported by ASTA Medica AG and then by VIATRIS GmbH & Co. KG, Frankfurt/Main, Germany, and in part by a grant from the German Federal Ministry of Health (BMG) and Federal Ministry of Science and Technology (BMBF).
Monitoring of the study was performed by Pharmalog, Munich. The authors thank colleagues from the Epidemiologic Study Group (Drs. S. Arlt, K. Henkel, A. Voigt, H. Tschampa, S. Hengst, D. Ruge, C. Werner, W. Murach, C. Riedemann, S. Kropp, B. Mollenhauer, M. Seipelt, K. Weidehass, C. Laske, C. Lassek, C. Jakobi, A. Schroeter, M. Bartl, K. Koertner), colleagues from the Department of Neuropathology (A. Giese, W. Schulz-Schaeffer), and all physicians reporting suspect cases to the German CJD Surveillance Unit for providing pertinent clinical, neuroradiologic, and biochemical data. They also thank Dr. H. Nowak for statistical support and Dr. N. van Ahsen and B. Ciesielcyk for laboratory support, and especially the colleagues who allowed them to include patients in the study.
Footnotes
-
See also page 679
- Received April 25, 2003.
- Accepted October 28, 2003.
References
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.