Mitochondrial disease (MD) is a clinically and genetically heterogenous group of
disorders. The dysfunction of the mitochondrial respiratory chain results in ATP shortage
and energy depletion resulting in multi system disease. Therapeutic interventions are
currently limited and focus on relieving symptoms, since no curative therapies exist. The
m.3243A>G related MD is presumably the most frequent MD affecting at least 1 in 5000
people, and the MIDD (Mitochondrial Inherited Diabetes and Deafness) phenotype is its
most frequent clinical phenotype. The role of nutrition, providing the essential
nutrients for mitochondria to produce energy, is an important disease modifier in MIDD.

A ketogenic diet (KD) is a very low-carbohydrate ( ±20 grams a day) high-fat diet that
shifts metabolism towards β-oxidation and ketone body production. In cell studies and
mice models for MD, ketone bodies or KD significantly affect different dimensions of
mitochondrial function. KD has been proven to effectively treat intractable epilepsy in
children with MD. The investigators recently performed a 12 week pilot study with KD in
20 adults with different types of mtDNA related MD. The KD improved glucose tolerance,
yielded less migraine, and alleviated gastrointestinal symptoms. Despite these positive
effects, it proved the KD is a difficult diet to follow and only 8/20 participants
finished the 12 week KD intervention. The individuals with m.3243A>G related MD had the
longest diet adherence. Because of the positive effects, over half of all participants
switched to the more feasible and less restrictive low carbohydrate (± 50-100 gram/dag),
high fat (LCHF) diet after completing the study. The positive effect of KD on the subtype
of m.3243A>G related MD was very recently confirmed in another study. The LCHF diet
however provides no ketone bodies and its effect in mitochondrial disease is not studied
properly yet.

In irritable bowel syndrome, specific dietary regimens as low-fermentable, oligo-, di-,
monosaccharides and polyols (FODMAPs) but also LCHF diets are considered therapeutic. A
diet low in fermentable carbohydrates reduces the fermentation by the gut microbiota and
is therefore thought to reduce bloating and other gastro intestinal complaints. Based on
the results of a previous study from the investigators and the effect of LCHF diets in
irritable bowel syndrome, the investigators hypothesize the same positive effect in the
cohort of m.3243A>G related MD in which gastro intestinal complaints are very common.

An adult diabetes team studied the gut microbiome and observed significant differences in
various species and pathways between the faecal stool samples of diabetes patients and
healthy controls, and that gut microbiome composition is associated with glycemic
control. The gut microbiome is a very promising study field. More insight in patients gut
microbiome and the way it can be altered with nutrition could lead to new treatment
options and symptom relieve which is very relevant to MD patients. Therefore it is
interesting to compare the gut microbiome of MIDD patients with already analysed stool
samples of healthy controls and diabetes patients, and study the effect of an LCHF diet
intervention.

The LCHF diet is an established treatment option in diabetes, with positive effects on
glucose homeostasis. The investigators have initially chosen the subgroup of m.3243A>G
related MIDD patients as they hypothesize that the LCHF could help slow down the
progression of diabetes severity and reduce the necessity for treatment with insulin in
these patients in particular. Since the team could not find enough participants with MIDD
they intent to include also participants with MD caused by the m3243A>G mutation but
without diabetes. These patients still have a lot of gastro intestinal problems as well
as fatigue and headaches, all relevant symptoms that possibly can be affected with the
low carbohydrate diet. So the primary outcomes as mentions in the title of the study:
effect on clinical symptoms and the gut microbiome are relevant for this patient
population aswell.

Studies in the heterogeneous patient population with MD are challenging. By choosing a
relatively frequent genotype the investigators can minimize this problem and study a
defined cohort. Additionally the investigators will use a cross over design ("patients
being their own controls") to further decrease the number of patients needed to be
included to achieve robust data.

Previous experience with trials in MD including patient consultations has led to an
insight that the Goal Attainment Scaling (GAS) is an important individually tailored
outcome for this patient group. With the GAS patients choose and score personal goals
that they want to achieve with the diet intervention.

Taken together, The investigators intend to perform a pilot study to investigate the
effect on clinical symptoms, as measured by GAS, as well as assessing gastrointestinal
complaints and the impact on the gut microbiome of a low carbohydrate, high fat diet
intervention. This study will focus on a cohort comprising adult patients exhibiting
m.3234A>G mutation-related MD. Additionally, within the subgroup of patients with
m.3234A>G-related MIDD, diabetes control will be also be evaluated.

1.1 Visits:

Visit baseline:

- Signing informed consent

- Final check inclusion and exclusion criteria

- Randomisation

- Secure baseline stool sample

- Check baseline food diary

- Vitals including blood pressure

- NMDAS , SF 12,

- GAS (Goal Attainment Scaling)

- Nutritional assessment: Weight, body composition (Multi frequency BIA SECA BCM 555)

- Online Questionnaires will be filled in Castor database during the visit under
supervision of the researcher to prevent missing data. Questionaires include: Gastro
intestinal complaints, + SF 12, CIS fatigue, diet tolerance, adverse effects and
medication

- Laboratory examinations (CK, HBA1c, lipid profile, kidney function, transaminases,
lactate, ketone bodies, metabolomics)

- Patient instructions on home measurements and if randomised in the intervention
group diet instructions including safety instructions.

Visit 12 weeks:

- NMDAS with SF 12

- Vitals including blood pressure

- GAS (Goal Attainment scaling)

- Nutritional assessment: Weight, body composition (Multi frequency BIA SECA BCM 555).

- Online Questionnaires will be filled in Castor during the visit under supervision of
the researcher to prevent missing data.

- Laboratory examinations (CK, HBA1c, lipid profile, kidney function, transaminases,
lactate, ketone bodies, metabolomics)

- Check on the home measurements and if randomised in the control group diet
instructions and if randomised in the intervention group instructions how to stop
the diet.

Visit 24 weeks:

- NMDAS with SF 12

- Vitals including blood pressure

- GAS (Goal Attainment scaling)

- Nutritional assessment: Weight, body composition (Multi frequency BIA SECA BCM 555).

- Online Questionnaires will be filled in Castor during the visit under supervision of
the researcher to prevent missing data.

- Laboratory examinations (CK, HBA1c, lipid profile, kidney function, transaminases,
lactate, ketone bodies, metabolomics)

- Check on the home measurements and diet instructions for after the study based on
patients wishes.

1.2 Home measurements:

- Diet adherence will be checked with a digital food diary 3 of days including one
weekend day. At baseline, 11 weeks and 23 weeks.

- Fecal sampling at baseline, 12 weeks and 24 weeks.

- Accelerometery: Activ8® has to be worn on the upper leg for 1 week at the time: At
baseline, 11 weeks and 23 weeks.

- Weight measurement once a week.

- Headache diary: Headaches need to be registered during the whole study

- Glucose monitoring will be done with sensor derived glucometrics at the last 10 days
of both the intervention and the control period. Two times (20 days) in total.
Exclusively conducted for MIDD patients. In MIDD patients without continuous glucose
monitoring in the usual care (this are the patients without intensive insulin
therapy) the sensor will be blinded.

1.3 Instructions:

- Instructions on when and how to fill in the online eating report.

- Instructions on how to wear the actometer at screening.

- Diet instructions at the visits in the intervention period and during weekly
telephone (or videocall) consultations during the first 4 weeks of the diet and
thereafter at week 8. Diet instructions by a specialist dietitian. The diet
instructions consist of a personalized diet plan and a patient brochure with links
to recipes accompanied by verbal instructions. Diet and, if necessary medication,
will be adjusted in the first few weeks based on glucose levels in blood,
complaints, practical questions and weight developments. This will be done in the
same manner as in regular patient care for adult diabetic patients who start with
the LCHF diet. Difference between this protocol and this study is that in the
protocol a diabetic nurse takes care of the medication changes and in this study Dr.
Rick Meijer is responsible, or in his absence another internist experienced in the
treatment of MIDD. Additional telephone or e-health consultations (mijnRadboud) can
be planned as frequent as necessary depending on the questions of the patient. A
specialised metabolic dietician is available 5 days a week. In the week after the
participant has to switch back to usual care there is a telephone meeting planned to
make sure the transition went safely and al information is clear for the
participant.

- Instructions on how to collect a faecal stool sample

- Instructions for the blinded continuous glucose monitoring. Participant have to
apply the sensor at home in week 11 and week 23. Exclusively for MIDD patients who
don't have continuous glucose monitoring already.

- Safety instructions on how to prevent hypoglycemia and how to act in case of a
hypoglycemic incident. These instructions will be personalized based on individual
diabetes medication and blood glucose levels, the protocol is used as guideline for
these personalized instruction. This protocol is usual care in diabetes patient care
and based on the national guidelines. The instructions included medication changes.
The safety instructions on how to adapt the diet to hypoglycaemic events are
included in the diet brochure.

Investigational treatment: Low carbohydrate high fat diet Low carbohydrate (50-100
grams/day). The LCHF restricts a wide range of carbohydrate rich foods including sweets,
sugar, grains, starchy vegetables, sweet legumes, fruit juices and sweetened dairy foods
and replacing them with non-starchy vegetables, red fruits, meat, poultry, seafood, eggs,
oil, nuts, and non sweetened high fat dairy foods such as cheese, butter, cream and Greek
10% fat yoghurt. Choosing cuts of meat higher in fat and fats, and the use of low carb
alternatives like low carb bread is encouraged. This diet is automatically high in
protein. We don't want the carbohydrate restriction to be stricter than 50 grams a day
because we want a feasible diet without de adverse events that a ketogenic diet (very
strict carbohydrate restriction of 20 grams a day) has. The low carbohydrate high fat
diet is part of the national diabetes guidelines and is frequently prescribed in
overweight diabetes type 2 patients with insulin resistance.

METHODS 2.1 Measurements

The questionnaires The Gastrointestinal complaints Questionnaire contains questions on
gastrointestinal complaints (7 point Likert scale) form a validated Dutch gastro
intestinal questionnaire. We used this questionnaire in studies with mitochondrial
disease before. For severity we use a Numerical Rating Scales (NRS) stool frequency,
Bristol stool chart and laxative use is being checked.

The general study questionnaire in the intervention phase there will be questions on
feasibility and diet tolerance, adverse events and questions on diabetes medication doses
(insulin use).

- Quality of life: SF-12 questionnaire The SF-12 Health Survey (see F1.SF-12-Dutch) is
a shorter version of the SF-36 Health Survey that uses 12 questions to measure
functional health and self-administered well-being. SF12 takes approximately three
minutes to complete and it is a practical, reliable and valid measure of physical
and mental health. It is useful in large population health surveys or for
applications that combine a generic and disease-specific health survey. SF 12 is
available in multiple languages, including Dutch. It covers the following eight
health domains: 1) Physical functioning 2) Limitations in usual role activities
because of physical health problems 3) Bodily pain 4) General health perceptions 5)
Vitality (energy / fatigue) 6) Limitations in social activities because of physical
or emotional problems 7) Limitations in usual role activities because of emotional
problems and 8) General mental health. Two subscales are derived from the SF-12: the
Physical Component Summary (PCS) and the Mental Component Summary (MCS). These
summary scales based on the SF-12 correlate very highly with the SF-36 version. The
SF-12 uses norm-based scoring, so comparisons can be made among the other generic
health surveys. The SF-12 and SF 36 are frequently used in mitochondrial disorders
and it is advised to always combine the NMDAS with one of these quality of life
questionnaires.

- Fatigue questionnaire: short checklist individual strength The Checklist Individual
Strength is a 20-item self-reported questionnaire that captures four domains of
fatigue: 1) subjective severity of fatigue 2) reduction in concentration, 3)
reduction in motivation and 4) reduction in activity. Respondents will rate the
extent to which each statement is true for them in the past two weeks on a
seven-point Likert scale. The scores are summed up accordingly. The higher the
score, the higher the fatigue level. A score between 27 and 35 suggests an increased
risk for fatigue and a score above 35 suggests severe fatigue. The questionnaire has
been used in various diseases, including mitochondrial diseases.

Questionnaires will be filled in digitally as eCRF in castor by the patients directly.

Nutritional Assessment: measurements will be performed according to our standard
operation procedures Weight (kg), Weight loss (%), Height(cm) BMI (kg/m2) and body
composition will be measured with bioimpedance analysis SECA BCM 555®. With the outcome
variables: Fat percentage. Fat free Mass (FFM) (kg) and FFM Index (kg/m2), Skeletal
muscle mass index (kg/m2) and phase angle(◦).

Blood test:

- creatinine kinase (CK),

- diabetes: HbA1c and home glucose monitoring (exclusively in MIDD patients)

- lipid profile (after overnight fast/or at least 4 hours of fasting): (triglycerides,
high-density lipoprotein (HDL), low-density lipoprotein (LDL), total cholesterol),

- kidney function (MDRD), transaminases (ALAT and ASAT), lactate

- Ketones (beta-hydroxybutyric acid and acetoactetate)

- metabolomics

Home blood test:

Diabetes control assessments will be exclusively conducted for MIDD patients. Glucose
monitoring can be done by blinded continues glucose monitors (Dexcom). Patients who are
accustomed to glucose monitoring at home as part of their routine care will not be
required to utilize the study-provided blinded monitor. Patients who use Insulin or
glibenclamide, gliclazide, glimepiride or tolbutamine will be provided with a not blinded
finger prick glucose meter to guarantee safety and to detect any low blood sugars in a
timely manner. Measuring blood glucose with this home device with a finger prick is
requested when participants experience complaints that may indicate low blood sugar, such
as extreme hunger, sweating, trembling, dizziness, palpitations, yawning, being confused
or feeling restless.

Three-day food record:

A digitally three-day food record is a self-reported account of all foods and beverages
consumed by a respondent over three days, two working days and one day of the weekend.
The data will be entered by the participant in the Eetmeter app.

The aim of the use of the food record is to finetune the diet and evaluate diet
adherence. It helps patients to get more insight in their carbohydrate intake and to
better manage the diet so they are free to use it more frequently (but for the study only
3 times 3 days are required). Intake will be automatically calculated in the eetmeter
app. Data on energy (kcal), carbohydrates (gram), protein (gram) fat (gram), fibre
(grams) and water (ml) intake will be registered.

Accelerometry:

The Activ8 is used to measure average metabolic equivalents(METS) over one week period.
Activ8 and the elaborate use instructions will be sent to patients' home (see
E4.Draaginstructie Activ8 actometer). Patients will apply the actometer and wear activ8
for one week during 24 hours per day. The device waterproof so the participants can also
wear it during shower or swimming. Actometer activ8 has to be worn on the upper leg for 1
week at the time in total 3 times: during the screening, week 6, and week 12. This is
used to calculate the energy expenditure for the individual diet advice[29] since the
amount of activity can be a confounder as well as an possible efficacy variable.
Actometry is considered the gold standard for activity measurements and it is a
non-invasive method[30].

NMDAS:

The Newcastle Mitochondrial Disease Scale for Adults (NMDAS) is a semi-quantitative
clinical rating scale designed specifically for all forms of mitochondrial disease. It
was developed and tested in Newcastle with the aim of providing a validated and
reproducible measurement of disease progression. It is now the most widely used outcome
measure in clinical trials for mitochondrial disease. The NMDAS scale considers many
parameters, including neurological disease, diabetes, cardiac involvement. The NMDAS need
to be filled during the screening and week 12 by the metabolic specialist directly in
castor.

Headache diary Frequency and severity of headages are scored with a headaches diary.
Headaches need to be registered only if occur. A NBS score for severity of the headache
and if patients took medication for the headache can be registered in the diary.

Other study parameters:

Patient characteristics: age, gender.

2.2 Randomisation, blinding and treatment allocation After signing informed consent
participants will be randomized in the intervention or control group with the castor
database. There is no blinding or treatment allocation in this study. Upon arrival on the
screening, the participants receive an ID code. The first participant will be number one
and so on.

2.3 Withdrawal of individual subjects Subjects can leave the study at any time for any
reason if they wish to do so, without affecting future management and treatment. The
investigator can decide to withdraw a subject from the study for urgent medical reasons.

2.4 Replacement of individual subjects after withdrawal Patients will not be replaced
after withdrawal. However, if patients do not meet the inclusion and exclusion criteria
based on the results from the first visit, then these patients will be replaced.

2.5 Follow-up of subjects withdrawn from treatment Patients will be asked if they are
willing to complete the final study visit after they stopped the diet. They are asked why
they stopped the diet but they are not obligated to answer this or come to the closing
visit.

3. STATISTICAL ANALYSIS

Descriptive analyses Statistical analysis will be performed using SPSS statistics version
24 (IBM). Baseline characteristics will be reported, at timepoint 0 weeks, as means ± SD
for normally distributed variables, as median and IQR for non-normally distributed
variables and as a percentage of the total population for the categorical variables.