Networks Integrating Anxiety and Metabolism in Human Anorexia Nervosa, A Randomised Controlled Trial on Physiological and Psychological Responses to Caloric and Non-Caloric Meals
Networks Integrating Anxiety and Metabolism in Human Anorexia Nervosa, A Randomised Controlled Trial on Physiological and Psychological Responses to Caloric and Non-Caloric Meals
Anorexia nervosa (AN) is known as the psychiatric disorder with the highest mortality rate as long-term calorie restriction causes medical complications and suicide risk. The ineffectiveness of pharmacological treatments increases interest in the etiopathogenesis of the disease. This study aims to examine the effects of caloric and non-caloric food intake on stress response, changes in brain network activity in AN patients and healthy volunteers, and examine whether these changes are related to caloric intake through neuroendocrine and psychological responses.
The study will include 30 female AN patients aged 13-18 who are newly diagnosed and have not received any treatment, and 30 healthy volunteers matched by age, sex, and education level. The patient group will be admitted to the clinic's open inpatient unit on the evening before the study to ensure they have fasted. A 12-hour fasting period starting from the day before the procedure will be mandatory.
In the morning, patients will undergo psychiatric assessments and scales. Psychometric measurements will be applied to determine the clinical characteristics of participants. Visual Analog Scale (VAS) will be applied. VAS is used to evaluate subjective states such as hunger, satiety, desire to eat, anxiety, fear of obesity, sadness, appetite, general mood, and motivation to participate in the study. Following this, the first resting-state functional magnetic resonance imaging (fMRI) scan will be performed for 8 minutes.
Subsequently, participants will be randomised with 1:1 allocation ratio to consume one of two jellies prepared by a dietitian: one containing 400 calories and the other calorie-free. Participants will not be forced to consume the food.
The ingredients and preparation methods were chosen to make the appearance and taste of the foods similar. Participants will not know which type of food (caloric/non-caloric) they are given. Participants who do not wish to consume the provided food will be excluded from the study. Participants will be told that the mixture they will be consuming will either be balanced in nutritional value and caloric, or harmless and calorie-free. They will be given fifteen minutes to consume the mixture, and the amount they consume will be recorded. Sixty minutes after the consumption of the mixture, participants will undergo a second fMRI scan to identify the neuronal network changes induced by this process. The same procedures will be applied to the healthy control group. Before and after imaging, blood samples will be collected from participants for biochemical analysis. Endocrine markers such as glucose, insulin, ghrelin, obestatin, PYY, leptin, and cortisol will be evaluated to assess hunger, satiety, and stress responses. Changes in these parameters will be used to evaluate participants' physiological responses to meals. A second meal will be offered three hours after the first meal, and the amount consumed will be recorded.
Four subgroups will be analysed in the study: AN patients who received calories, AN patients who did not receive calories, healthy controls who received calories, and healthy controls who did not receive calories. Outcome data will be collected before and after meal intake and the outcomes will be compared between the four groups.
The study aims to investigate the physiological, psychological and behavioural responses to caloric and non-caloric meals, and to compare the responses in patients and non-patients. It is hypothesized that the stress response observed in AN patients is not solely related to food intake but that the calorie content of the ingested food will cause additional activation in biologically stress-related networks, with corresponding effects on experience. The hypothesis that the differences observed between patients and controls are due to biological differences in AN patients will be tested. This study aims to shed light on the etiopathogenesis of AN and contribute to the development of new strategies for the treatment and management of AN.
Methods
Design The study is designed as a single-center, parallel-group RCT with two study arms: (1) calory intake in meal-1 and (2) no calory intake in meal-1. Participants and healthy controls are individually randomized to either one of the two arms, both delivered at the eating disorder clinic of the Ege University Hospital, Faculty of Medicine.
Primary outcome measure is brain patterns observed with magnetic resonance imaging and secondary outcomes include endocrine markers in blood, cardiovascular responses, self-reported symptom severity, and subsequent food consumption.
The study is planned in accordance with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) and will be analysed and reported in accordance with the recommendations in consolidated standards of reporting trials (CONSORT), with special care taken to meet extended recommendations for both trials with patient reported outcomes and non-pharmacological trials.
Protocol Design and Conduct To ensure transparency and integrity throughout the study, all changes or deviations from the protocol will be thoroughly documented and reported. Any modifications to the protocol, including changes to study objectives, design, participant population, sample size, procedures, or significant administrative aspects, will necessitate a formal protocol amendment. Similarly, any alterations that could potentially impact participant safety, benefits, or risks will be carefully evaluated and documented. Proposed amendments will be reviewed and approved by the study team and submitted for approval to the national ethical review board before implementation. Additionally, all relevant study personnel will be promptly informed of these changes to ensure compliance and consistency across the study.
Recruitment The study will begin to recruit participants in February 2025 and the last measurements for the primary manuscript are expected to be collected by the end of 2026.
Recruitment, interventions and data collection will be performed in the specialized eating disorder clinic of the Faculty of Medicine at Ege University Hospital in Izmir (population approximately 4.5 million), Turkey. Before participants are recruited, we will ensure the implementation of Good Clinical Practice guidelines, the establishment of a functioning clinical research infrastructure and the training of study staff.
Participants will be recruited through four possible pathways: (1.) AN patients will be recruited from Ege University Faculty of Medicine Hospital, by physicians meeting these patients. Healthy controls will be recruited through (2.) announcements on Ege university campuses, (3.) local hospitals, and (4.) online platforms such as Instagram.
Individuals who agree to participate will be contacted by study personnel for further information about the study over the telephone. At that time an initial assessment for eligibility will be performed. For further eligibility assessments, please see the next heading. For adolescents that cannot be reached over the telephone, parents or legal guardians will be contacted. Potentially eligible participants will be scheduled to a study physician for further eligibility assessment and written informed consent.
Eligibility Female individuals of 13-18 years of age, who are currently patients at the eating disorder clinic and both newly diagnosed with AN according to the Diagnostic and Statistical Manual of mental disorders 5th edition (DSM-5) and treatment naïve will be eligible. Healthy control participants will be matched on age-, sex-, and education level with the AN-participants.
The clinical diagnosis/diagnoses for participants will be validated/ruled out by the Structured Clinical Interview for DSM-5 (SCID-5). Please see "Baseline measures for eligibility" for instrument details. The presence of other exclusion criteria will also be further explored at the clinical visit.
The exclusion criteria are: (1.) Having one or several chronic medical conditions that may interfere with or hinder participation, or affect study outcomes, (2.) Use of medications that affect hormone levels or eating behaviors, (3.) Having one or several severe psychiatric disorders other than AN (e.g., schizophrenia, major depressive disorder, bipolar disorder), (4.) Current drug abuse, pregnancy or potential pregnancy (for AN patients these conditions are excluded as a part of routine care. For healthy controls no objective testing will be performed), (5.) High risk of refeeding syndrome or severe complications related to eating disorders, (6.) Inability to comply with the nutritional program of the study, (7.) Allergies to foods included in the provided meal content, (8.) Cognitive impairments preventing comprehension of the study or the informed consent process, (9.) Contraindications for MRI (e.g., pacemaker, prosthesis, claustrophobia), and (10), fasting blood sugar above 6.0 mmol/l on the morning of the study, as this indicates either non compliance with overnight fasting or pre-diabetes.
Additional exclusion criterion for healthy controls: (1.) Having one or several eating disorders, current or past.
Unexpected findings of biochemical anomalies at T0 (please see "Assessments" below) will not be considered exclusion criteria and medical conditions will be referred for appropriate treatment.
Participant inclusion in the study will be determined by a study clinician based on the outlined eligibility criteria. In cases of uncertainty, the principal investigator will make the final decision. Importantly, to maintain impartiality and avoid selective recruitment bias, enrolling clinicians will be masked to future treatment allocations, as randomization will occur at a later point in time. Participants will receive both oral and written information about the study and will provide written informed consent before inclusion. For participants below 18 years of age, written informed consent will also be obtained from parents or legal guardians. Participation in the study is voluntary, and participants may withdraw at any time without any negative consequences or affected subsequent care. Participants may also choose to withdraw any personal data already collected.
Randomization When an eligible participant has been recruited and consented, they will be scheduled for randomization at a later timepoint. AN Participants will be individually randomized to one of the two study arms with a 1:1 allocation ratio, by a computer-generated allocation sequence using permuted blocks. The block sizes will vary randomly between two and four.
Randomization lists will have been pre-prepared by a separate researcher who is not involved in participant recruitment. Sealed envelopes will contain the study-ID for each participant and the meal assigned to that individual (caloric or non-caloric). When a participant is ready to enter the study, a researcher will draw an envelope and prepare the assigned meal for that participant in the evening, for consumption in the next morning.
Participants with AN will be admitted to the clinic's open inpatient unit on the evening prior to study participation to ensure a 12-hour overnight fast. On the following day, they will receive the specified interventions and undergo the outlined assessments. Once an AN participant has been randomized and completed the study, a matched healthy control participant will be recruited. Recruitment of healthy control participants will be conducted by study personnel who are masked to the randomization outcomes of the corresponding AN participant. This approach ensures that the allocation of both AN and healthy control participants remains concealed from recruiting researchers and those conducting assessments. Healthy control participants will undergo the same eligibility assessments as AN participants, be assigned their own study ID, and share the same randomization outcome as their corresponding AN participant. However, unlike AN participants, healthy control participants will not be admitted overnight.
Codes will be used to increase information confidentiality and participant anonymity, and participant allocation will be performed on the basis of these codes. Randomization will continue until n = 30 AN-patients and n = 30 healthy controls have completed the study as per-protocol participants.
The randomization results will not be revealed to anyone other than the individual preparing the assigned meal. The randomization step will thus be separated from the patient enrolment process and all the other processes of the study. Through that participants, study personnel and clinicians delivering any co-interventions will be masked to participants' allocation, and so will outcome-assessors and statistical analysts.
Intervention protocol Depending on their randomization result participants will be asked to consume one of two jellies, both composed by a dietitian and prepared by study personnel: one containing 400 kilocalories and the other calorie-free. Both participants and the person handing out the jelly will be unaware of its caloric content. Participants will be given 15 minutes to consume the jelly, without forcing them to do so, and the amount consumed will be recorded.
The caloric food used in the study will be prepared using jelly powder, water, powdered sugar, cinnamon, ginger, turmeric, and lemon zest. To make the mixture, 100 ml of hot water will be poured over the plain jelly powder and stirred until fully dissolved. Eighty grams of powdered sugar will then be added and stirred until completely integrated. For flavor enhancement, 1 teaspoon each of cinnamon, ginger, and turmeric, along with the zest of half a lemon, will be incorporated. Afterward, 100 ml of cold water will be added, and the mixture will be stirred thoroughly. The prepared mixture will be poured into single-serving containers, allowed to cool at room temperature, and then refrigerated for 2-3 hours to solidify.
The calorie-free mixture will be made with 200 ml water, agar-agar, a natural sweetener, cinnamon, ginger, turmeric, lemon zest, and lemon juice. These ingredients will be combined and brought to a boil on the stove. Once boiling, the mixture will be stirred continuously for 5 minutes until it begins to thicken. The heat will then be turned off, and the mixture will cool for 15 minutes. It will be transferred into containers and refrigerated for 3 hours to set.
The ingredients and preparation methods were carefully chosen to ensure that both mixtures have a similar appearance and taste.
Three hours after consuming the first mixture (meal-1), participants will be offered a second meal (meal-2). This meal will always contain calories and consist of a nutritionally balanced, standard oatmeal dish with a distinct taste from the first meal. Participants will be allowed to eat freely, up to a maximum of 500 kcal, to minimize the risk of medical complications such as refeeding syndrome in participants with AN. Participants will have 45 minutes to complete the meal, and the total amount consumed will be recorded.
Apart from meal-1 and meal-2, any oral intake apart from water will be prohibited during the time of the study.
Adherence to study protocol Protocol non-adherence is defined as: (1.) A participant withdraws their consent, choosing to discontinue the study from the time of randomization to the end of T2 data collection, (2.) A participant is discovered to have received the wrong type of meal in meal-1, (3.) A participant consumes ≤ 30 % of the food provided in meal-1, (4.) A participant eats or drinks anything other than the provided meals and water between T0 and T2 data collection, (5.) A participant refuses or is unable to undergo MRI-scanning at T2, (6.) A participant is unable to adhere to the study schedule, and (7.) An adverse event occurs that renders continued participation inappropriate, please see "Adverse events and security plan" below for details.
All outcome measures detailed under "Assessments" will be collected from all participants. For the primary analysis, all data from participants adhering to protocol will be included. For details see the "Statistical Methods" section below.
Adverse events and security plan Adverse events (AEs) are defined as any undesirable experiences requiring healthcare that occur to a participant during the study, regardless of whether they are related to the study or not. All AEs reported spontaneously by participants or their relatives, or observed by investigators, study personnel, or clinicians, will be documented in the electronic case report form. Both AEs and serious adverse events (SAEs) - including those resulting in death, being life-threatening, requiring hospitalization, or involving other significant medical concerns - will be promptly reported to the principal investigator. SAEs that are assessed as potentially related to the study will be reported to the accredited ethics committee that approved the protocol within seven days of the responsible researcher becoming aware of the event.
Based on previous studies and unpublished preliminary data, significant deterioration of symptoms during the study is not anticipated. However, for safety purposes, self-report questionnaires collected throughout the study will be reviewed by study personnel. Patients are expected to become anxious when asked to eat a meal and regular meal-related symptom fluctuations will not be considered as adverse events. Any other observed worsening of symptoms or adverse reactions will be addressed in discussions with the participant and, when appropriate, their parents or legal guardians.
As the interventions and assessments are provided as part of the regular healthcare system, all participants are covered under standard patient injury insurance. External study monitoring will be performed to ensure that all data storage, handling, and analysis procedures comply with ethical guidelines, regulatory requirements, quality standards, and protocol.
Assessments Pre- and post-meal self-report scales will be administered via a secure online platform accessible only to the research team. Pseudonymized data from these scales, as well as all other assessments, will be stored on a secure server with regular security backups. Biological specimens collected during the study will be coded and stored in coolers in a certified biobank for batchwise analysis after the dataset is complete. Physical records will be stored in locked cabinets with restricted access. The pseudonymization key will be securely stored in a location that is separate to data.
Participants travel expenses and food as well as beverage costs will be covered and this amounts to approximately 20 US dollars. Those who fail to attend scheduled appointments as a part of the study will be reminded via telephone or email. Parents and/or legal guardians may also come to be contacted.
Eligibility measures The medical records of all participants will be analyzed to identify comorbidities and other exclusion criteria. The biochemical results that will be available for AN-patients include lipid profile (total cholesterol, HDL, LDL, and triglyceride levels), blood glucose level, insulin level, TSH, T4, FSH, LH, estradiol, prolactin, and CRP. These parameters are routinely checked on all AN-patients. No additional blood tests will be performed to identify exclusion criteria and for healthy controls these blood tests will not be performed.
Clinician interviews will be performed to verify the clinical AN diagnosis in patients and to rule out any comorbidity in AN and healthy control participants. The SCID-5 will be used to complement the clinical interview, and it will be performed by a well-trained psychiatrist used to applying the instrument in practice. The SCID-5 is a reliable and valid tool for diagnosing psychiatric disorders and its psychometric properties in Turkish are good.
Assessments for description of sample Body weighth and height will be measured at T-1 for a calculation of body mass index. At the same time the participants' sociodemographic background will be assessed using a brief self-made questionnaire. Additionally, at T-1 the following self-rating scales will be collected for screening purposes and description of sample: (1.) The Autism Spectrum Screening Questionnaire (ASSQ), (2.) the Children's Depression Inventory (CDI), (3.) the Childhood Anxiety Screening Scale (ÇATÖ), (4.) the Eating Attitudes Test (YTT), and (5.) the Global Assessment Scale (GAS).
Outcome measures Outcome measures will be assessed at timepoints T0, T1, T2, and T3, with some measures collected only at specific timepoints. Details on this for each outcome measure are provided under their respective subheadings and summarized in Table 1. If a participant misses a scheduled data collection timepoint, the following steps will be taken in sequence: (1) if feasible and meaningful, the data will be collected as soon as the issue is identified; (2) if not, the data point may be recorded as missing, and the participant will continue in the study; or (3) if the participant is deemed no longer relevant to the study, they will be excluded from further follow-up and per-protocol analysis.
Self-report At T0, T1, T2, and T3 psychometric evaluations will be performed on Visual Analog Scales (VAS). The following questions will be used at the appropriate timepoints: (1.) How anxious are you about consuming the meal? Response options: I feel no anxiety - I feel very anxious, (2.) How often do you feel anxious about eating unfamiliar foods? Response options: I do not feel anxious - I feel very anxious, (3.) How hungry do you feel? Response options: I feel full - I feel very hungry. (4.) How worried are you about the possibility of gaining weight from consuming the meal? Response options: I feel no anxiety - I feel very anxious, (5.) How much do you fear becoming obese? Response options: I do not fear - I fear a lot, (6.). How much does the thought of becoming obese affect your eating habits? Response options: It does not affect at all - It affects a lot, (7.) How overweight do you feel? Response options: I feel very thin - I feel very overweight.
Blood sampling At T0 and T2 blood samples will be collected from all participants. At T0 participants' fasting glucose will be measured. Neuroendocrine measurements on blood samples taken at T0 will include ghrelin, obestatin, peptide YY (PYY), leptin, and cortisol hormones. These analyses will be conducted using available kits. At T2 the same neuroendocrine parameters will be analyzed in blood.
Blood pressure and pulse To evaluate cardiovascular responses blood pressure and heart rate measurements will be performed using an automatic blood pressure monitor and a pulse oximeter (both Sun Tech Medical, Oxfordshire, England) at T0 and T2.
Magnetic Resonance Imaging Resting-state MRI scans will be performed at T0 and T2. MRI examinations will be conducted using a 3.0 Tesla MRI scanner (Siemens Magnetom Verio, Numaris/4, Syngo MR B17, Erlangen, Germany) equipped with a 12-channel head coil. First, structural images will be obtained with axial TSE T2-weighted, coronal 3D-SPACE FLAIR, and 3D-MPRAGE sequences. Structural images will be reviewed by an on-site radiologist to identify any brain anomalies or pathologies that would render a potential participant ineligible (e.g., structural abnormalities, lesions). If no pathological findings are detected, imaging will continue with a fMRI sequence. Including 9 minutes of resting-state fMRI where participants will keep their eyes open and look at a fixation cross on screen.
Meal-1 and meal-2 food consumption Although not strictly considered an outcome the amount of mixture (in grams) consumed in the meal-1 will be recorded and included as a control-variable in the analysis. The amount of food consumed in meal-2 will also be recorded by weighing the plate before and after the meal.
Statistical methods Analysis plan Analysis will be conducted once data collection is complete. No interim analyses will be performed. Initially, the quality and completeness of the data will be assessed. If available, missing values will be entered using information from the individual's original data file. The dataset will be examined for outliers and inconsistencies. For categorical variables, any responses outside the defined categories will be coded as missing. For continuous variables, values outside the acceptable ranges will also be coded as missing.
The analyses will be carried out by the investigators and a biostatistician using the latest version of SPSS Statistics (IBM Corp., Armonk, NY, USA), R (R-core team, Vienna, Austria), or equivalent statistical software. For the analysis of fMRI data additional software tools described below will be used. Data analysts will be blinded to the participants' allocation. Significance testing will be two-tailed, and statistical uncertainties will be reported as 95% confidence intervals. All tests will be conducted with a significance level of 0.05.
Descriptive statistics Baseline data and descriptive statistics will be reported using standard measures in accordance with CONSORT guidelines and the baseline characteristics of the included participants will be reported per randomization group and AN status in a baseline table. The following baseline characteristics will be reported: age (years), AN subtype (by categories) comorbidity (% yes, by diagnostic categories), smoking (% yes, frequency), country of birth (% Turkey, % first generation immigrant and % second generation immigrant-with at least one parent born outside of Turkey). Sex will not be reported as all participants will be of female sex.
Patient flow diagram The flow of participants will be illustrated in a flow diagram according to CONSORT recommendations.
Primary analysis For the statistical analysis of functional neuroimaging data we will use MATLAB and Statistical Parametric Mapping (SPM12) software to assess brain activity patterns and differences across groups. Connectivity analysis will be performed using the CONN Toolbox to examine the relationships between different brain regions and networks. Neuroimaging data will first undergo preprocessing to ensure the data is suitable for further statistical analysis.
The preprocessing steps that will be performed include, briefly, (1.) Realignment: This is the process of correcting head motion of the participants, (2.) Slice-Time Correction: This involves correcting for functional differences between slices acquired every 3 seconds, (3.) Coregistration: Anatomical and functional images are overlaid to determine the anatomical location of activation, (4.) Spatial Normalization: This process accounts for individual brain shapes by normalizing anatomical MRI images. The template developed by the Montreal Neurological Institute (MNI) will be used, and (5.) Smoothing: This is used to improve image quality and will be performed using a 3D Gaussian kernel (FWHM = 8mm).
After the preprocessing steps, the data of all participants will be registered to the common Harvard-Oxford Cortical Atlas space. Participant and group-level regression analyses will be performed in this atlas space. Functional connectivity analyses will be performed between regions of the components of the Salience Network (SN), including the anterior insula, dorsal anterior cingulate cortex, temporoparietal junction (TPJ), and hypothalamus. Changes in the topology of networks formed by these four regions, in addition to their functional connectivity with each other, will be analyzed using Graph Theory. The changes in the connectivity features of these regions before and after caloric/non-caloric intake will be evaluated using both nodal and global metrics. Significant results obtained will be recalculated using masks transferred back to the participant space. Beta values of the regression analysis will be controlled for Type-1 errors with false discovery rate and reported accordingly.
The independent variables of the regression analyses will include (1.) Type of meal (caloric or non-caloric), (2.) AN status (AN or healthy control), (3.) AN subtype (restrictive or bulimic type), (4.) Body Mass Index, and (5.) Duration of AN illness.
Secondary analyses Secondary outcome variables will be modelled using similar regression models as used in the primary analysis. In these models, time (T0/T1/T2) and the allocation by time interaction will be included as fixed effects when analyzing variables that are measured at more than two timepoints. Additional parametric tests will be performed as long as the assumptions for them are met. Depending on the type of data, the following statistical methods will be used: Independent t-test, Pearson r, and ANOVA. Where necessary, corresponding non-parametric alternatives will be used.
Dose-response analyses Additional dose-response analyses will be performed by controlling the previously specified regression models for the amount of mixture consumed in meal-1.
Sample size calculation A power calculation was performed using G*Power software for a t-test with linear regression (two independent groups), determining that 27 participants per group would be sufficient. To account for potential dropouts or unforeseen issues, 30 participants will be recruited for each group to ensure robust statistical analysis.
Inclusion criteria:
Healthy control participants will be matched on age-, sex-, and education level with the AN-participants. The clinical diagnosis/diagnoses for participants will be validated/ruled out by the Structured Clinical Interview for DSM-5 (SCID-5). The presence of other exclusion criteria will also be further explored at the clinical visit.
Exclusion criteria:
Additional exclusion criterion for healthy controls: (1.) Having one or several eating disorders, current or past.
Participant inclusion in the study will be determined by a study clinician based on the outlined eligibility criteria. In cases of uncertainty, the principal investigator will make the final decision. Importantly, to maintain impartiality and avoid selective recruitment bias, enrolling clinicians will be masked to future treatment allocations, as randomization will occur at a later point in time.
olof.lagerlof@umu.se+46 90-785 00 00
erik.ekback@umu.se+46 660-890 00