Study Design
This is a single-center, randomized, controlled, three-armed, parallel-group study conducted at the NeuroCure Clinical Research Center at Charité - Universitätsmedizin Berlin. Recruitment started in April 2017 (ClinicalTrials.gov identifier: NCT03508414). Patients are recruited from all over Germany. Recruitment strategies include specific study calls on the website of the German Multiple Sclerosis Society and distribution of study flyers in specialized neurological practices as well as information events and lectures for patients. We intend to randomize 111 RRMS patients to one of three 18-month dietary interventions. The dietary interventions are 1) KD with a carbohydrate intake restricted to 20-40 g/day, 2) a FD with a 7 day fast every 6 months and 14 hours daily intermittent fasting in between, and 3) a SD that is predominantly vegetarian and fat-modified as recommended by the German Nutrition Society. The institutional review board of Charité - Universitätsmedizin Berlin approved the study and written informed consent is obtained from all participants prior to study entry (Additional files 1 and 2). The study is conducted in accordance with the Declaration of Helsinki in its currently applicable version, the guidelines of the International Conference on Harmonization of Good Clinical Practice (ICH-GCP) and applicable German laws.
Participants
Patient information and informed consent have been prepared in accordance with the guidelines of the institutional review board of Charité - Universitätsmedizin Berlin. Potential participants receive both forms at least 24 hours before a consultation with a study physician who personally explains all study procedures. If he or she is willing to participate and had sufficient time to ask questions, written informed consent is given. Afterwards, inclusion and exclusion criteria are assessed. Main inclusion criteria are a definite diagnosis of RRMS according to the 2017 McDonald criteria (46), a stable immunomodulatory therapy or no disease-modifying therapy for at least 6 months, an EDSS score below 4.5 (47) and disease activity within the last two years before study entry. This is defined as at least one new lesion on brain MRI and/or at least one relapse. Participants are recruited in cohorts of 15-25 patients.
Complete list of inclusion criteria
- Relapsing-remitting MS
- 18-65 years of age
- EDSS < 4.5
- Stable immunomodulatory therapy or no disease-modifying therapy ≥ 6 months before study entry
- ≥ 1 relapse or ≥ 1 new T2 lesion or ≥ 1 contrast-enhancing lesion on MRI within the last two years
- Agreement that possible incidental findings will be communicated
- Body mass index between 19 and 45 kg/m2
- Ability to give verbal and written consent
- Health insurance
Complete list of exclusion criteria
- Start or change of immunomodulatory therapy < 6 months before or during the study
- Relapse or cortisone treatment within 30 days before study entry
- Clinically relevant metabolic, progressive or malignant diseases
- Intake of >1g/day omega-3 fatty acid supplements
- Significant cognitive-cooperative impairment
- Insulin-dependent diabetes mellitus (type I)
- Participation in another interventional study
- Weight loss diet or loss of more than 5 kg within 2 months before study entry
- Insufficient mental ability for cooperation
- Eating disorders
- Kidney stones
- Therapy with oral anticoagulants
- Pregnancy and breastfeeding
- Suspected lack of compliance
- Smokers (> 5 cigarettes per day)
- Known alcohol and drug abuse
- Inability to give informed consent or apply to the study protocol
- Contraindications for MRI such as metallic implants, cardiac pacemakers, claustrophobia
Retention of patients is promoted by close and frequent contact to their nutritional counselor and study physician via telephone, email and study visits. In case of relevant issues, visits outside the study schedule are offered.
Discontinuation criteria are withdrawal of consent, subsequent occurrence of an exclusion criterion (e.g. change of immunomodulatory therapy), lack of compliance and medical reasons for stopping the intervention.
For all three groups, compliance is defined as attending at least 8 of 10 group sessions and each of 6 study visits. Food records measure compliance to the prescribed intervention. In the KD group, majority of blood ketone measurements should be ≥ 0.5 mmol/L. For the FD group, there are regular additional meetings during each of the three fasting periods. These meetings have to be attended enabling physician and counselor to evaluate compliance.
We offer patients who meet a discontinuation criterion to attend the remaining study visits for follow up outside the study protocol. We assess all outcomes except for exploratory blood parameters (autophagy, oxidative stress) in these patients.
Randomization
Eligibility of patients will be determined at the screening visit by a trained physician. Afterwards, randomization is done in three strata to distribute possible confounding factors equally to the dietary interventions. Strata are sex (man or woman), MS medication (yes or no) and lesion load (<15 or ≥15 lesions) in the baseline brain MRI scan. We use a stratified block randomization with variable block length. This will ensure a homogeneous distribution of interventions over these strata. An external statistician who is not involved in the study carries out the randomization according to predefined randomization lists.
Dietary interventions
Nutritional counseling for all three interventions takes place in small groups within 10 sessions over 18 months (5 sessions month 1-3, 5 sessions month 4-18). The group setting facilitates exchange between patients, management of dietary challenges, monitoring of compliance and recording of adverse events. Between group sessions, patients have the possibility to contact their nutritional counselor via email or telephone at any time.
KD: During nutritional counseling patients are instructed to start by limiting carbohydrate intake to just 20 g per day for four weeks in order to establish ketosis. Then, patients increase their carbohydrate intake by 5 g each week until they reach their individual maximum (approx. 40 g) to maintain stable ketosis. All carbohydrates relevant for elevating blood glucose are limited to 40 - 50 g/d. In addition, glycemic index and glycemic load of carbohydrates have to be below 50 and 6, respectively. This ketogenic diet is equivalent to a traditional ketogenic diet, but with a liberalized macronutrient composition of 70-80% fats, 15-20% proteins and 5-10% carbohydrates (compared to a traditional ketogenic diet with 90% fat, 6% proteins and 4% carbohydrates). To aid in the adjustment and determine their individual carbohydrate intake limits, patients receive a hand-held ketone-meter to measure and record blood concentrations of ketone body ß-hydroxybutyrate at regular intervals. Values should be between 0.5-3.0 mmol/L.
FD: Patients in the FD group fast for 7 days every six months. During this intensive fasting episode, patients only consume vegetable juices and vegetable broth yielding a daily calorie intake between 200 and 350 kcal. Tea and water is ad libitum to assure sufficient fluid intake. Two days before and three days after fasting patients eat a low-calorie vegetarian diet for preparation and aftercare. The fasting is initialized by an intestinal clearing with laxatives (e.g. Glauber´s Salt by FX-Passage). Outside these intensive fasting episodes, patients intermittently fast for 14 hours daily on seven days of the week in order to maintain fasting effects. For compliance reasons, participants may take one “cheat day” per week at which they follow the fasting rules less strictly. During intermittent fasting, patients are counseled to eat a diet according to the recommendations of the DGE (equivalent to the SD group). Patients additionally attend group sessions every other day during fasting. A physician and a dietician with expertise on fasting monitor these sessions. During the first intensive fasting episode, these sessions are every other day. In the subsequent fasting episodes, the meetings are less frequent (3-4 meetings).
SD: Patients are counseled to adhere to a healthy, non-calorie-restricted diet according to the recommendations of the DGE. This diet is predominantly vegetarian with reduced consumption of meat, animal fats, eggs and egg products. Low-fat milk and dairy products are recommended to provide calcium. Dietary recommendations are designed to modify the omega-6- to omega-3 fatty acid ratio to 5:1.
Adverse events
There are no major risks for patients participating in this study. Minor adverse events of a ketogenic diet and fasting can be headaches, feelings of hunger, fatigue, irritability and dizziness. These side effects are transient and will stop after a few days. Patients will be asked for tolerability of the interventions and any adverse events will be recorded. We do not expect serious adverse events due to our dietary interventions.
Outcome parameters
Our primary endpoint is the number of new lesions on cranial T2-weighted MRI after 18 months compared to baseline. Secondary MRI endpoint is brain atrophy determined by percent brain volume change. All MRI scans will be performed in a 3-Tesla MRI scanner (Tim Trio, Siemens, Erlangen, Germany) and will be assessed by an experienced evaluator who is blinded to both clinical data and interventional allocation.
Additional secondary endpoints are annualized relapse rate, neurological-functional disability progression (EDSS, Multiple Sclerosis Functional Composite), cognition (Symbol Digit Modalities Test), fatigue (Fatigue Severity Scale), depression (Beck Depression Inventory II), muscle strength (handgrip dynamometer), walking endurance (six min walk test) and quality of life (MSQoL-54). In a subset of patients, we will evaluate glucose variability from continuous glucose measurements over 14 days (FreeStyle Libre Sensors, Abbott). All endpoints will be assessed at baseline, 9 and 18 months by trained and experienced personnel.
For safety monitoring body weight and composition, dietary intake and vital signs are recorded at regular intervals as well as routine laboratory tests of blood count, kidney and liver parameters in peripheral venous blood. In addition, we want to investigate metabolic, hormonal and immunological effects of the diets. For this, peripheral blood mononuclear cells are isolated for further analysis. In addition, markers of oxidative stress, especially reaction products of anti-oxidative enzymes, are of interest. Polyamine level and cell metabolites will be measured as secondary markers of nutrient availability and autophagy activation in the cell. Stool samples are collected for 16s rRNA sequencing. For a detailed overview of assessments and endpoints see table 1.
We will inform participants about any conspicuous findings and refer them to their general practitioner for further treatment. After completion of the study, participants have the possibility to be referred to appropriate outpatient nutritional counselling.
Power calculation
Based on the results of other MS studies in our research center in comparable patient cohorts, an average of 3 new T2 lesions after 18 months is expected in the SD group. Sample size was calculated with Wilcoxon (Mann-Whitney) rank sum test and significance level was adjusted (Bonferroni) for the two comparisons (KD vs. SD, FD vs. SD). With a sample size of 33 patients per group and a two-sided significance level of 0.025, there is a power of 80% to detect a probability of 0.72, that an observation from group 1 is smaller than group 2. This corresponds to a standardized effect size of 0.83 and an expected mean difference in the number of new T2 lesions in each of the intervention groups of 0.5 compared to the SD group after 18 months (Query Advisor 7.0). We expect a dropout rate of approximately 10% and therefore plan to enroll 37 patients per group, in total 111 patients.
Data management
Each participant will receive a unique identifier upon study entry. This identifier will be used for all data documentation to assure participant’s confidentiality. Data will be gathered in source documents and then transferred into paper case report forms. Later all data will be digitized and stored in a central database. To improve accuracy of data entry, entries will be verified for proper format and expected range as well as double-checked. Overall data quality will be assured by independent monitoring throughout the study. However, due to the minimal risks of our dietary interventions, a data monitoring committee is deemed unnecessary. Data will be stored for 10 years after study completion and then be deleted. Any modification to the current study protocol will be submitted to the institutional review board, all trial participants, and trial investigators.
We only obtain consent for usage of data and samples for the research question described in this protocol. Thus, we do not intend to use participant data or biological samples in ancillary studies.
Results will be personally explained to all study participants, presented on national and international conferences, published in peer-reviewed journals, and disseminated to neurologists, and the medical laity. We will comply with the official eligibility guidelines for authorship for all publications and do not intend to use professional writers.
Data analysis
Confirmatory analysis will be conducted based on the intention-to-treat (ITT) population. The aim is to show that KD and FD are superior to SD meaning that the number of new T2 lesions after 18 months of dietary intervention adjusted for the baseline value is lower in the KD and FD group than in the SD group. Though sample size calculation was conservatively based on the Mann-Whitney U test, the final analysis of the primary endpoint will be based on the non-parametric ANCOVA-type (Analysis of Covariance) analysis on ranks with treatment group as fixed factor and adjustment for baseline number of T2 lesions. The resulting coefficients will be reported with 95% confidence intervals and their corresponding p-values compared to a two-sided Bonferroni adjusted significance level of 0,025 (48).
We do not plan any interim analyses. In case of more than 5% missings for the primary endpoint, we plan to replace missing values in the outcome parameter using multiple imputation (49) under missing at random assumption in addition to the complete case analyses.
As a sensitivity analysis, the primary endpoint will be analysed in the per-protocol population. This population includes all randomized patients who meet study eligibility criteria and fulfil all compliance criteria throughout the study.
We intend to do two subgroup analyses; men vs. women and MS medication vs. no MS medication by adding another main effect for sex or medication, respectively, as well as an interaction term. These subgroup analyses will be analyzed in an exploratory manner to generate possible hypotheses for follow up studies without adjustment for multiple comparisons. For these, p-values may not be interpreted as confirmative.
Safety analysis will include calculation of frequencies and rates of adverse and serious adverse events within 18 months of the interventions.
Secondary endpoints will be analyzed in an explorative manner and descriptives will be given using mean and standard deviation for sufficiently normal distributed metric variables, median with limits of the interquartile range [IQR: 25th and 75th percentile] for skewed metric or ordinal data, as well as absolute and relative frequencies for count data.