Effects of diazoxide in multiple sclerosis

Patients.

Men and women aged 18–55 years with relapsing-remitting MS (RRMS, McDonald criteria 2010⁷) and an Expanded Disability Status Scale (EDSS) score of 0–5.0 were eligible for this study. The inclusion criteria required at least 1 relapse in the previous 2 years or the presence of at least 1 Gad+ lesion in the previous 1 year. The exclusion criteria were (1) patients who are candidates for treatment with drugs modifying the course of the disease according to the criteria of the regulatory agencies in each country, unless the patient refused to initiate such therapy or decided to postpone the start of this therapy; (2) patients who experienced a relapse in the 30-day period prior to the baseline visit; (3) patients on treatment with diazoxide; (4) patients with a medical condition such as hypotension, insulinoma, hyperuricemia, or diabetes defined by American Diabetes Association criteria; (5) patients with other conditions, including drug abuse, liver transplantation, inability to provide informed consent, or inability to perform all the procedures in the clinical trial; (6) patients with contraindications for MRI studies; (7) patients with contraindications for treatment with diazoxide, excipients, or gadolinium-based agents; (8) patients on corticosteroid therapy in the last month; (9) patients on interferon β or glatiramer acetate therapy in the past 3 months; (10) patients on natalizumab or fingolimod therapy in the past 6 months; (11) patients treated with chemotherapy (mitoxantrone, azathioprine, cyclophosphamide, cladribine, methotrexate) or who entered previous trials of treatments under development; and (12) patients with a positive pregnancy test, those currently breast-feeding, or women of childbearing potential not using highly effective methods of contraception.

Standard protocol approvals, registrations, and patient consents.

The trial was approved by the local institutional review board at each center and by the Spanish and German medical agencies. Patients were included after providing their informed consent.

Study design.

This was a double-blind, placebo-controlled, multicenter phase 2a study to determine the safety of diazoxide and to search for evidence of efficacy of 2 doses of diazoxide in patients with RRMS (ClinicalTrials.gov: NCT01428726; EUDRA-CT: 2010-023048-34). The trial was conducted at 16 sites (13 in Spain and 3 in Germany) in accordance with the Helsinki Declaration⁸ and adhering to the International Conference on Harmonization Guidelines for Good Clinical Practice.⁹ All investigators and patients were blinded to the treatment and to the blood glucose and diazoxide levels, unless they were alerted for safety reasons. Diazoxide and placebo were prepared as identical tablets with no differences in taste or appearance.

One hundred two patients were randomly assigned to 3 groups: diazoxide 0.3 mg/d, diazoxide 4 mg/d, or placebo (ratio 1:1:1) administered orally once daily. Block randomization was performed using 2 blocks of 6 elements with random sequences of the 3 arms in each center using a randomization seed in order to ensure its reproducibility and using SPSS v17 software and macro Generation of Random Sequences (!RNDSEQ).

Doses were selected on the basis of preclinical studies in an animal model of MS^4,5 in order to maintain the doses below those at which diazoxide induces hyperglycemia or hypotension (100 mg/kg/d). Patients received diazoxide in 1 daily oral dose of 0.3 mg or 4 mg (total dose) from day 1 to the end of week 24. Study visits took place at screening, at baseline, on day 1 (randomization), and every 4 weeks until the end of the 24-week treatment period or drug discontinuation. Treatment allocation was concealed without breaking the blinding until the end of the study. At the end of the follow-up, the patients could enroll in an additional 24-week observation period.

Confirmed relapse was defined as the occurrence of new symptoms or a worsening of previously stable or improving symptoms and signs lasting more than 24 hours, not associated with fever, and accompanied by an increase of at least half a point on the EDSS. When warranted, relapses were managed by the treating physician according to a standardized scheme: IV administration of up to 1,000 mg of methylprednisolone per day for 3–5 days. For the duration of the methylprednisolone therapy, glycemia was monitored daily using a glucose finger prick test. Diazoxide treatment was maintained during the relapses and the study schedule was not changed (the visit and MRI schedule remained unaltered) unless patients needed to use prohibited concomitant treatments. EDSS was assessed by an independent neurologist not involved in patient care at screening, at baseline, every 4 weeks, and at unscheduled visits in the case of MS relapse.

MRI.

MRI scans were performed at baseline and every 4 weeks until the end of the study (week 24). Patients enrolled in the follow-up study were subjected to an additional scan at week 48. Patients received a double dose of the contrast agent (0.2 mmol/kg), and the same agent was used for all patients at the same center, with a latency of 10 minutes. New and enhancing lesions were identified by certified radiologists, and T1 and T2 lesion volumes were determined by trained technicians using the Jim software (Xinapse Limited Systems, Aldwincle, UK). The normalized brain volume at baseline and the percentage brain volume change (PBVC) between baseline and month 24 were calculated using the Structural Image Evaluation using Normalization of Atrophy program.¹⁰ Black holes were defined as areas of unequivocal low signal intensity compared with normal-appearing white matter that did not show contrast uptake and were concordant with a hyperintense lesion seen on T2-weighted imaging. Finally, exploratory tissue-type segmentation was used to calculate separate estimates of gray matter, white matter, and CSF volumes with Statistical Parametric Mapping (SPM5). MRI analysis was performed at the MRI unit of the “Centro de Esclerosis Múltiple de Cataluña” (Cemcat), Hospital Vall d'Hebron, Barcelona, Spain.

Study endpoints.

The primary efficacy endpoint was the cumulative number of new Gad+ lesions on T1-weighted sequences from week 4–24. We selected this endpoint based on the effects of diazoxide on microglia activation and closing of the BBB and because this endpoint has been validated in many previous short-duration trials (6 months). Secondary MRI endpoints were cumulative number of new/enlarged lesions on T2-weighted sequences for all MRIs; cumulative number of new/increased lesions on T2-weighted sequences in the 6 months after starting therapy (compared with the baseline MRI); cumulative number of combined unique active lesions (CUALs), addition of new or enlarged lesions on T2-weighted sequences that do not enhance with gadolinium and new Gad+ lesions for all MRIs; cumulative number of CUALs, addition of new or enlarged lesions on T2-weighted sequences that do not enhance with gadolinium and new Gad+ lesions in the 6 months after starting therapy (compared with the baseline MRI); number of patients without Gad+ lesions in T1-weighted sequences in the 6 months after starting therapy; and PBVC. Secondary clinical endpoints were relapse-free status, relapse rate, number of relapses requiring corticosteroid treatment, time to first relapse during the trial, change in EDSS, and quality of life assessed with the Short Form 36 Health Survey (SF-36). Secondary safety endpoints were incidence, nature, and severity of adverse events (AEs) during the 6 months of therapy up until 15 days after the last dose of diazoxide or placebo, including control of glucose levels, glycated hemoglobin, blood pressure, and presence of hirsutism.

Statistical analysis.

The primary alternate hypothesis was that the cumulative number of new T1 Gad+ lesions observed in all MRIs would be distinct in patients treated with diazoxide (0.3 or 4 mg/d) and those who received placebo. The null hypothesis was tested in a negative binomial regression model, as we considered this to best explain the distribution of new Gad+ lesions.¹¹ The analysis was planned for an intention-to-treat (ITT) population (any patient who was randomized, received treatment, and had at least 1 postbaseline efficacy variable completed) and a per-protocol (PP) population (any patient who was randomized and received the protocol-required treatment and processing). In the PP population, no data input methods were used. In the ITT population, when primary outcome observations were missing in the study visits, the patient was assigned the median of the number of new lesions for the remaining visits.

The univariate analysis of categorical variables involved the χ² or Fisher exact test, and continuous variables were assessed with a t test, a Mann-Whitney test, or a Wilcoxon test, as appropriate. Differences in the number of new T2 lesions and CUALs were analyzed using the negative binomial distribution, as described previously.¹¹ Paired samples with continuous variables were analyzed using a t test or Wilcoxon signed-rank test. The time-to-event analysis was achieved with a Cox analysis. Multivariate analysis was performed with linear generalized models or generalized estimating equations with the evolution between visits as an outcome measure. p Values ≤0.05 were considered significant in all tests, which were carried out with the SAS statistical package v9.2.

The underlying assumption for calculating the sample size was a minimum effect between groups of 2 lesions and a common SD of 2.55.¹² We assumed a power of 80% and an α error of 0.025. Sample size estimation was based on the t test. We took into account that endpoint assessment would be performed after a 6-month follow-up, so we estimated a loss rate of 10%. Finally, the study sample size was calculated with GRANMO 7.10 software, which indicated that 35 patients were needed in each group (a total of 105 patients) to detect a treatment effect ≥2 lesions with a common SD of 2.55 and a loss rate of 10%.

Primary research questions and classification of evidence.

The aim of this study was to test the safety of diazoxide and to search for signs of efficacy in patients with RRMS. This study provides Class I evidence that diazoxide is safe and well-tolerated in patients with MS. It does not decrease the presence of new lesions on MRI and it slows the progression of brain atrophy.