Fasting, ketogenic, and anti-inflammatory diets in multiple sclerosis: a randomized controlled trial with 18-month follow-up.

The researchers compared three dietary interventions head-to-head over 18 months:

**Standard healthy diet (SD):** A mainly vegetarian diet following German Nutrition Society recommendations, with an anti-inflammatory focus by targeting an omega-6 to omega-3 fatty acid ratio of 5:1. No calorie restriction. No intermittent fasting. This served as the active control group.

**Fasting diet (FD):** Participants completed a 7-day Buchinger-style fast (very low calorie, typically 200–300 kcal/day from juice and broth) every 6 months (so two total fasts over 18 months). Between fasts, they practiced 14:10 time-restricted eating (14 hours fasting, 10-hour eating window) on 6 days per week. Dietary recommendations between fasts matched the SD group.

**Ketogenic diet (KD):** Participants restricted carbohydrate intake to 20–40 grams per day and increased fat intake to 70–80% of total energy, with emphasis on plant-based fats. They could gradually increase carbs up to their individual limit (maximum 50g/day) to maintain stable ketosis, defined as blood beta-hydroxybutyrate (BHB) ≥ 0.5 mmol/L. Participants self-monitored with hand-held ketone meters.

The primary outcome was the number of new T2-hyperintense lesions on brain MRI after 18 months. Secondary outcomes included brain atrophy, serum neurofilament light chain (sNfL — a marker of nerve damage), cognitive function (Symbol Digit Modalities Test), depressive symptoms, and cardiometabolic risk markers (BMI, abdominal fat, blood lipids, blood pressure).

**Sample size:** 105 individuals with relapsing-remitting multiple sclerosis (RRMS) were enrolled; 81 completed the study (77% completion rate).

**Population:** Adults with a definite diagnosis of RRMS according to 2017 McDonald criteria, with stable disease-modifying therapy (or no therapy) for at least 6 months before enrolment, and recent disease activity (at least one new MRI lesion or one clinical relapse within the prior 2 years).

**Disability level:** Expanded Disability Status Scale (EDSS) score < 4.5 (meaning they could walk without aid and were not severely disabled).

**Setting:** Single center at the Neuroscience Clinical Research Center of Charité – Universitätsmedizin Berlin, Germany. Recruitment from April 2017 to October 2021.

**Key exclusion criteria:** Starting or changing disease-modifying therapy within 6 months before study start or during the study.

**Primary outcome — New T2 lesions:** Brain MRI scans at 3-Tesla (Siemens MAGNETOM Trio Tim and Prisma scanners) at baseline, 9 months, and 18 months. Lesions were segmented by two experienced, blinded MRI technicians.

**Brain atrophy:** Percentage of brain volume change (PBVC) at 18 months vs. baseline.

**Serum neurofilament light chain (sNfL):** Measured retrospectively after study completion using the Quanterix NFLight® Assay at an accredited laboratory. sNfL is a protein released when nerve axons are damaged — higher levels indicate more active neurodegeneration.

**Cognitive function:** Symbol Digit Modalities Test (SDMT), a standard measure of information processing speed where participants match numbers to symbols within 90 seconds. Higher scores = better cognitive function.

**Depressive symptoms:** Measured via a validated scale (specific scale not detailed in abstract, but likely Beck Depression Inventory or similar).

**Cardiometabolic markers:** Body mass index (BMI), abdominal fat (likely waist circumference or DXA), blood lipids (cholesterol, triglycerides), adipokines (hormones from fat tissue), and blood pressure.

**Ketosis monitoring (KD group only):** Blood beta-hydroxybutyrate (BHB) measured with hand-held meters (GlucoMen LX plus), with target ≥ 0.5 mmol/L.

**Study design:** This was a single-center, randomized, controlled, parallel-group trial with three arms. Participants were randomly assigned 1:1:1 to standard diet, fasting diet, or ketogenic diet.

**Randomization:** Participants were stratified by three factors before randomization: (1) disease-modifying therapy use (yes/no), (2) sex (male/female), and (3) T2 lesion load (low ≤15 vs. high >15). Block randomization was performed by an external statistician not involved in study procedures. This stratification ensures that these known confounders were balanced across groups.

**Blinding:** This was an open-label study — participants and study personnel delivering the dietary interventions obviously knew which diet they were following (you cannot blind someone to whether they are fasting or eating a ketogenic diet). However, outcome assessors for MRI and EDSS (disability scoring) were trained and blinded to group assignment. This is a critical design feature: while participants know their diet, the people measuring the primary outcome (MRI lesions) do not, reducing bias in the most important endpoint.

**Duration:** 18 months total, with measurements at baseline, 9 months, and 18 months. This is unusually long for a dietary intervention trial in MS — most previous studies lasted 3–6 months.

**Dietary intervention delivery:** Participants attended 10 group sessions over 18 months. The FD group had additional sessions every second day during their 7-day fasts. From March 2020 to December 2021, sessions were conducted virtually due to COVID-19 restrictions, with enhanced phone and online support to maintain adherence.

**Statistical approach:** The primary analysis compared the number of new T2 lesions after 18 months in FD and KD groups vs. SD group. Secondary endpoints were analyzed exploratorily (meaning no correction for multiple comparisons — this increases the risk of false-positive findings). The study used intention-to-treat principles where possible.

**What this design can prove:** A randomized controlled trial with blinded outcome assessment can establish whether dietary interventions cause changes in MRI lesions compared to a control diet. The long duration (18 months) is a strength for detecting slow disease progression. Stratified randomization balances key confounders.

**What this design cannot prove:** Without blinding of participants, there is potential for placebo effects on subjective outcomes (depression, cognition). The single-center design limits generalizability. The study cannot determine whether effects are specific to the dietary interventions or result from the increased attention and support that dietary intervention groups receive (the SD group also received group sessions, partially controlling for this). The exploratory analysis of secondary outcomes means positive findings should be considered hypothesis-generating, not confirmatory.

**Major methodological weaknesses:**

**High dropout rate:** 24 of 105 participants (23%) did not complete the study. If dropouts differed systematically between groups, this could bias results.

**COVID-19 disruptions:** The study was conducted partly during the pandemic, with group sessions moving to virtual format. This may have affected adherence and outcomes differently across groups.

**No correction for multiple comparisons:** Secondary outcomes were analyzed exploratorily, increasing the risk of Type I error (false positives).

**Single center:** Results may not generalize to other populations or healthcare settings.

**Active control group:** The SD group was not a true "no intervention" control — they received dietary guidance and group support. This makes it harder to detect differences between diets but is ethically necessary for an 18-month study.

**Primary outcome — New T2 lesions at 18 months:**

Standard diet (SD): 0 new lesions (median, with range 0 to -1 — the negative value suggests some lesions may have resolved)

Fasting diet (FD): 0 new lesions (range 2 to 0)

Ketogenic diet (KD): 0 new lesions (range 2 to 0)

No statistically significant difference between any groups. The primary hypothesis — that fasting or ketogenic diets would reduce new brain lesions compared to standard diet — was not supported.

**Secondary outcomes (exploratory analysis — interpret with caution):**

**Serum neurofilament light chain (sNfL):** In the FD group, sNfL concentrations were lower at 9 months compared to baseline (mean decrease of -1.94 pg/mL, p = 0.042). This effect was not sustained at 18 months. No significant changes in SD or KD groups.

**Cognitive function (Symbol Digit Modalities Test):** In the KD group, SDMT scores improved by +3.7 points at 18 months compared to baseline (p = 0.020). No significant changes in SD or FD groups.

**Depressive symptoms:** In the FD group, depressive symptoms showed a trend toward improvement at 18 months (p = 0.079) — this did not reach conventional statistical significance (p < 0.05).

**Cardiometabolic markers:** BMI, abdominal fat, blood lipids, adipokines, and blood pressure improved in all three groups at 9 months, but the patterns differed between groups. Some improvements were partially associated with clinical outcomes in the FD and KD groups (specific numbers not provided in the abstract).

**Brain atrophy:** Percentage of brain volume change did not differ significantly between groups (specific numbers not provided in the abstract).

**New MRI lesions:** The median number of new lesions was zero in all groups. This means that over 18 months, most participants in all three diet groups did not develop new visible brain lesions. This could reflect effective disease-modifying therapy in the study population, or it could mean that all three diets were equally protective — but without a true no-intervention control, we cannot distinguish these possibilities.

**sNfL reduction (FD group):** A decrease of ~2 pg/mL at 9 months. To put this in context, sNfL levels in healthy adults are typically around 5–10 pg/mL, while in active MS they can range from 10–50+ pg/mL. A 2 pg/mL reduction is modest — roughly equivalent to a 10–20% decrease from baseline in someone with moderately elevated levels. This effect was temporary (gone by 18 months).

**Cognitive improvement (KD group):** A +3.7 point improvement on the SDMT. The SDMT has a range of 0–110, with typical scores in healthy adults around 50–60. A 3.7-point improvement is small but clinically meaningful — roughly equivalent to reversing 1–2 years of age-related cognitive decline. For comparison, some MS medications show improvements of 2–4 points on this test.

**Depressive symptoms (FD group):** The p-value of 0.079 means there is a 7.9% probability that the observed improvement occurred by chance alone. This is not statistically significant by conventional standards (p < 0.05), so we cannot confidently say the fasting diet improved depression.

**What the authors acknowledge:**

The study was conducted partly during COVID-19, requiring protocol modifications (virtual group sessions) that may have affected adherence and outcomes.

The primary endpoint showed no difference between groups, limiting conclusions about disease-modifying effects.

Secondary outcomes were analyzed exploratorily without correction for multiple comparisons.

The study was not powered for subgroup analyses.

**What a critical reader would note:**

**No true control group:** The SD group received active dietary guidance and group support, making it an "active comparator" rather than a true control. This makes it harder to detect differences and means we cannot know if all three diets were equally effective (or equally ineffective) compared to no dietary intervention.

**High dropout rate (23%):** If participants who dropped out had worse outcomes or poorer adherence, the results could be biased toward showing benefit.

**Open-label design:** Participants knew their diet assignment, which could influence subjective outcomes (cognition, depression) through expectation effects.

**Single center in Germany:** Results may not generalize to other countries, healthcare systems, or MS populations with different dietary habits.

**Stable disease-modifying therapy:** Most participants were on stable DMT, which is standard of care. The dietary effects may differ in people not on DMT, but this subgroup was too small to analyze.

**No blinding of dietitians or participants:** While MRI assessors were blinded, the people delivering and receiving the intervention were not, introducing potential bias in adherence and reporting.

**Industry funding:** Not explicitly stated in the abstract, but the study was conducted at a university medical center. No obvious pharmaceutical industry involvement, but the ketone meter manufacturer (GlucoMen) may have provided devices.

For someone running their own n=1 experiment:

**What to test:**

**Option A (ketogenic diet):** Restrict carbohydrates to 20–40g/day, increase fat to 70–80% of calories (focus on plant-based fats like avocado, olive oil, nuts, seeds). Use blood ketone meter to confirm ketosis (BHB ≥ 0.5 mmol/L). This is the intervention that showed cognitive improvement in the study.

**Option B (intermittent fasting with periodic extended fasts):** Practice 14:10 time-restricted eating (14-hour fast, 10-hour eating window) on 6 days per week, plus a 7-day water-only or very-low-calorie fast (200–300 kcal/day from juice/broth) every 6 months. This is the intervention that showed reduced nerve damage markers at 9 months.

**Option C (anti-inflammatory diet):** Mainly vegetarian diet with omega-6 to omega-3 ratio of 5:1. This is the control diet that also showed cardiometabolic improvements.

**Minimum meaningful duration:**

For cognitive effects (ketogenic diet): At least 18 months. The cognitive improvement in the study was only seen at 18 months, not at 9 months.

For nerve damage markers (fasting diet): At least 9 months, but the effect was temporary — you would need to test at multiple time points.

For MRI lesions: 18 months minimum, and even then, the study found no effect. Do not expect to see changes in brain lesions within weeks or months.

**What to measure (specific metrics):**

**Primary outcome:** Track your MS relapses (date, duration, severity) and any new symptoms. This is more practical than MRI for n=1.

**Cognitive function:** Take the Symbol Digit Modalities Test (available online) monthly. A meaningful improvement would be +3–4 points sustained over 6+ months.

**Nerve damage:** sNfL is not available for home testing, but you could ask your neurologist to order it at baseline and 9–12 months.

**Ketosis (if doing KD):** Blood BHB ≥ 0.5 mmol/L, measured with a hand-held meter. Test daily for the first 2 weeks, then weekly.

**Cardiometabolic health:** Weight, waist circumference, blood pressure, fasting blood glucose, and blood lipids (cholesterol, triglycerides) at baseline and every 3 months.

**Depressive symptoms:** Use a validated scale like the Beck Depression Inventory or PHQ-9 monthly.

**Adherence:** Keep a daily food log or use a tracking app. For KD, track daily carb intake and weekly ketone levels.

**Key confounds to control for:**

**Disease-modifying therapy:** Do not change your DMT during the experiment. The study required stability for 6 months before and during the intervention.

**Seasonal effects:** MS symptoms can vary with season (worse in summer heat). Run your experiment for at least 12 months to capture a full seasonal cycle.

**Stress and sleep:** Both affect MS symptoms and cognitive function. Track sleep quality (hours, interruptions) and perceived stress (weekly 1–10 scale).

**Exercise:** Physical activity improves MS outcomes. Keep your exercise routine constant