A randomized, double-blind, futility clinical trial of creatine and minocycline in early Parkinson disease

Background: Creatine and minocycline were prioritized for testing in Phase II clinical trials based on a systematic evaluation of potentially disease modifying compounds for Parkinson disease (PD). Objective: To test whether creatine and minocycline alter the course of early PD relative to a predetermined futility threshold for progression of PD in a randomized, double-blind, Phase II futility clinical trial. Agents that do not perform better than the futility threshold are rejected as futile and are not considered for further study. Methods: Participants had a diagnosis of PD within 5 years, but did not require medications for the management of symptoms. The primary outcome was the change in the total Unified Parkinson’s Disease Rating Scale (UPDRS) score from baseline to either the time when there was sufficient disability to warrant symptomatic therapy for PD or 12 months, whichever came first. Subjects were randomized 1:1:1 to receive creatine 10 g/day, minocycline 200 mg/day, or matching placebo. The futility threshold was set as a 30% reduction in UPDRS progression based on the placebo/tocopherol arm of the Deprenyl And Tocopherol Antioxidative Therapy Of Parkinsonism (DATATOP) trial. p Values ≤ 0.1 indicate futility. Results: Two hundred subjects were randomized to the three groups. Neither creatine (p = 0.96) nor minocycline (p = 0.66) could be rejected as futile based on the DATATOP futility threshold. The rate of progression for the calibration placebo group fell outside the 95% CI for the DATATOP historical control. In a sensitivity analysis, based on the threshold derived from the calibration placebo group, again neither drug could be rejected as futile. Tolerability was 91% in the creatine group and 77% in the minocycline group. Common adverse events included upper respiratory symptoms (26%), joint pain (19%), and nausea (17%). Conclusions: Both creatine and minocycline should be considered for definitive Phase III trials to determine if they alter the long term progression of Parkinson disease (PD). Additional factors must be weighed before selecting agents for Phase III trials, including safety, tolerability, activity, cost, and availability of these two agents in comparison with other agents currently in development for PD. NEUROLOGY 2006;66:664–671 Creatine and minocycline were prioritized for clinical testing in Parkinson disease (PD) based on a systematic review of potentially neuroprotective compounds. This review considered the quality and consistency of preclinical and clinical data.1 Both drugs are neuroprotective in preclinical models of PD, are bioavailable with oral administration,2 and penetrate the blood– brain barrier.3-5 Creatine monohydrate is a dietary supplement that has generally been used for improving performance in athletes. There have been no major safety or tolerability problems with oral supplementation of creatine in doses as high as 20 g per day for short periods.2 Creatine plays an important role in mitochondrial energy production. Creatine is converted to phosphocreatine and can transfer a phosphoryl group to ADP to make ATP, thereby buffering intracellular energy stores. There is evidence of mitochondrial dysfunction in PD, with deficits in complex I activity in platelets of patients with early PD6,7 and in postmortem substantia nigra pars compacta (SNpc) tissue of more advanced patients.8 Oral supplementation with creatine has been shown to protect against 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) induced dopamine depletion in mice9,10 and is protective in transgenic rodent models of Huntington disease (HD)11,12 and amyotrophic lateral sclerosis (ALS).13,14 Minocycline is a semi-synthetic second generation tetracycline used for treating a variety of infections. It has been used chronically in diverse conditions such as acne15 and rheumatoid arthritis.16 Minocycline exerts anti-inflammatory effects independent of antimicrobial effects. The loss of dopaminergic neuEditorial, see page 626 See also pages 628 and 660 *The NINDS NET-PD Investigators are listed in the Appendix. Sponsored by the NIH (National Institute of Neurological Disorders and Stroke), U01NS043127, U01NS043128, and U10NS44415 through 44555. Disclosure: The authors report no conflicts of interest. Received April 4, 2005. Accepted in final form November 30, 2005. Address correspondence and reprint requests to Dr. Bernard Ravina, 1351 Mt. Hope Ave., Suite 220, Rochester, NY 14610; e-mail: [email protected] 664 Copyright © 2006 by AAN Enterprises, Inc. rons in the SNpc is associated with an inflammatory response mediated by glial activation. Dopaminergic neurons are highly susceptible to microglial mediated injury which may promote neurodegeneration.17 Minocycline was shown to be protective in preclinical studies in the MPTP model of PD,18,19 and transgenic models of HD20 and ALS,14,20,21 although there are reports of negative results in an HD22 transgenic model and exacerbation of MPTP toxicity.23 Based on this information we conducted a randomized, double-blind, futility clinical trial of creatine and minocycline, in parallel, in early PD. Our goal was to test whether creatine and minocycline had the potential to alter the short-term course of early PD relative to a predetermined threshold for progression of PD. Futility trials are a type of Phase II clinical trial that have been used successfully in cancer research for many years.24,25 Futility trials are intended to eliminate agents that show low potential for further development.24,26 Agents found to be futile in comparison to the threshold would not be considered for further clinical testing. Agents that are not found to be futile do not automatically advance to Phase III clinical trials, but receive consideration for further testing based on the overall profile including safety, tolerability, and activity. A discussion of the methods, use, and interpretation of futility trials is presented in a companion article.27 Methods. Organization. This multicenter clinical trial was organized by the Clinical Trials Coordination Center (CTCC) at the University of Rochester, the Department of Biostatistics, Bioinformatics, and Epidemiology at the Medical University of South Carolina, and National Institute of Neurologic Disorders and Stroke. National Institute of Neurologic Disorders and Stroke sponsored the trial. The Steering Committee developed the protocol and consent forms with the participating sites and guided the implementation of the trial. The protocol and consent were approved by a National Institute of Neurologic Disorders and Stroke appointed Oversight Board (OSB), an independent Data Safety Monitoring Board (DSMB), and the institutional review boards of each of the participating sites. The DSMB monitored the safety, data integrity, and progress of the trial. Participants. Participants were men and women age 30 and over who had a diagnosis of PD but did not require medications for the management of their symptoms. Two of three cardinal manifestations of PD (tremor, rigidity, and bradykinesia) were required; these findings had to be asymmetric. The diagnosis of PD must have been made within 5 years of randomization. Women of childbearing potential were required to use adequate birth control and have a negative pregnancy test at baseline. Subjects were excluded if they had any secondary causes of parkinsonism, such as drug induced parkinsonism or structural lesions; had atypical parkinsonian syndromes; gait freezing or impairment in postural reflexes; had prior stereotaxic surgery for PD; used creatine, minocycline, or any investigational agent within 90 days prior to randomization; had known hypersensitivity to creatine or minocycline; used CoQ10 in doses greater than 300 mg 90 days prior to randomization; or had any clinically significant medical condition that could interfere with the subject’s ability to safely participate in the study or to be followed. Study design and randomization. A single arm futility study design was used to assess each of the study drugs, creatine and minocycline. We included a placebo arm for calibration, to verify and update the historical control assumptions used in sample size estimation.28 Eligible subjects were randomly assigned in a 1:1:1 fashion to receive 1) 10 g/day of creatine monohydrate and placebo minocycline, 2) placebo creatine monohydrate and 200 mg/day of minocycline, or 3) placebo creatine monohydrate and placebo minocycline. The primary futility analysis was at 12 months of follow-up, but each subject was followed for 18 months for additional safety information. Subjects and investigators were kept blinded to treatment group. On January 25, 2005, the 12 month database was locked and the primary analyses were conducted. Study intervention. Creatine was administered as 5 g sachets mixed with 8 ounces of liquid taken twice a day and minocycline was administered as 100 mg capsules taken twice a day. Both were taken with meals. Creatine and matching placebo were provided by The Avicena Group, Inc., and minocycline and matching placebo were purchased from Medicis Pharmaceuticals Corp. Outcome measures. The primary, prespecified outcome measure was the change in the total Unified PD Rating Scale (UPDRS) score from baseline to either the time at which there was sufficient disability to warrant symptomatic therapy for PD or 12 months, whichever came first. Disability was assessed by the site investigator, based on impairment in ambulation, activities of daily living, and occupational status. Investigators were trained on this endpoint using case vignettes and retrospectively each decision to initiate therapy was reviewed by members of the Steering Committee and the DSMB. The mean change in total UPDRS for each treatment group was compared to a prespecified futility threshold of a 30% reduction in the historically derived change in the total UPDRS, which was based on the placebo arm of a previous clinical trial.26 Tolerability was defined as the proportion of subjects taking study drug for the full 12 months. All severe adverse events (SAEs) were reviewed by the study medical monitor and an independent medical monitor. Both the site investigator and the medical monitors assessed the potential relationship between SAEs and study drug. Study procedures. At the screening visit, the purpose and potential risks and benefits were explained to potential subjects and each subject gave written informed consent. Subjects then had a baseline medical history, physical examination, and underwent the UPDRS. Blood was obtained for serum chemistry and complete blood count. Participants were re-evaluated at 1, 3, 6, 9, and 12 months (±6 days) after the baseline visit using the battery of clinical scales and blood was drawn again at 6 and 12 months. Sample size and statistical analysis. The sample size estimation was based on data from patients on placebo/tocopherol participating in the Deprenyl and Tocopherol Antioxidant Therapy of Parkinsonism trial (DATATOP),29 a large cohort of newly diagnosed patients with PD similar to our planned study population. The DATATOP study met the Pocock criteria for the use of historical controls.30 The observed mean change from baseline of total UPDRS in placebo/tocopherol patients (13 months ± 30 days) was 10.65 (SD 10.4). The threshold value was defined as 30% less progression on the total UPDRS than the 10.65 unit change in DATATOP, or 7.46. A sample size of 58 per group provides power greater than 85% to reject the null hypothesis of non futility if in fact the true mean total UPDRS worsening is greater than the threshold of 7.46 at the design alternative of 10.65. As with most clinical studies, a certain degree of noncompliance (including subject withdrawal or lost-to-follow-up) was expected. Assuming the noncompliance rate to be minimal at 5%, the required sample size was increased to 65 per treatment arm to account for the noncompliance in the intent-to-treat analysis.31 For each study arm, the set of statistical hypotheses tested was as follows: H0: Li 7.46 vs Ha, Li > 7.46 (=10.65), where Li is the mean change score (total UPDRS at 12 months or at the time of initiation of symptomatic therapy – total UPDRS at baseline) for the treatment arm i and 7.46 is the maximum mean increase (worsening) in the score between baseline and 12 months sufficient to warrant further evaluation of the drug in the Phase III trial. The hypothesis was tested with a one-sample t test at one-sided alpha level of 0.10. If the null hypothesis was rejected (p 0.1) then the drug would be considered futile for further testing in a Phase III trial. A secondary analysis of the primary outcome was planned, if the mean change in the total UPDRS score observed in the calibration placebo group falls outside of the 95% CI of the historical control mean change score of 10.65 (±1.02). The historical rate derived from DATATOP would be updated by incorporating the information from the calibration placebo group using Bayesian methods to derive a posterior mean. The futility threshold would be recomputed as 70% of this posterior mean and a one-sample t test would be performed for each active treatment arm. Analysis of the primary outcome was conducted under the March (1 of 2) 2006 NEUROLOGY 66 665 Figure 1. Flow diagram from screening to study completion (time until 12 months or need for symptomatic therapy, whatever comes first). Values were im­ puted due to missing visits for one creat­ ine and one placebo patient and for those terminating prior to completion listed above. No imputation was required for the two minocycline and one creatine pa­ tients who dropped out after initiation of symptomatic therapy. Reasons for drop­ ping out included the following: malaise, death (creatine), cardiac arrhythmia, lack of efficacy (minocycline), and colon cancer, depression (placebo). Of the 434 prescreened, 77 declined (major reasons being travel requirements, family’s ad­ vice, or doctor’s advice), 76 were excluded (major reasons being exclusionary medications or PD was too advanced), and the remaining 81 were identified as potential subjects (via chart review) but never screened. Seven signed inform consent but did not enroll (see text). intent-to-treat principle where all randomized subjects are in­ cluded in the analyses. For the small proportion of subjects who were lost to follow-up, we imputed their UPDRS change scores using the worst change score observed within their respective treatment groups. Exploratory analyses included multiple imputa­ tion to account for missing values and a sensitivity analysis in which we used a 30% reduction from the observed calibration placebo value to recompute the futility threshold value. No twosample comparisons were made between placebo and treatment groups as the study was not designed or powered for this type of analysis. A two sample comparison would be underpowered com­ pared to the planned futility analysis. Results. Subjects enrolled. Between June and Novem­ ber of 2003, 434 potential participants were identified and prescreened for eligibility. Of these, 207 (48%) participants met study entry criteria and consented to further screen­ ing at the 45 participating sites. Three potential study subjects were found to be ineligible and four did not enroll for unknown reasons. The remaining 200 eligible subjects were randomized to one of three treatment groups (figure 1). The treatment groups were similar at baseline on de­ mographic variables and total UPDRS and UPDRS subscores (table 1). Futility. Compliance with study visits was high and only two patients in the creatine group and no patients in the minocycline group had missing values requiring impu­ tation. The mean change (SD) in total UPDRS from either baseline to 12 months or the time at which symptomatic therapy was needed was 5.6 (8.69) for the creatine group and 7.09 (8.71) for the minocycline group (table 2) (figure 2). The observed progression in both the creatine and mi­ nocycline groups did not exceed the predetermined futility threshold. Therefore, the null hypothesis that the means were less than or equal to the threshold value of 7.46 (30% less than the 10.65 DATATOP historical rate of progres­ sion) could not be rejected for creatine (p = 0.96) or mino­ cycline (p = 0.63). Creatine and minocycline could not be rejected as futile using this analysis and therefore met the criteria for consideration for further clinical testing. Using multiple imputation instead of worst change score for the group to account for missing observations yielded similar