Study design
The randomized, open, cross-over clinical trial will be conducted at the Careggi University Hospital, Florence, Italy. A cross-over design will be implemented to allow comparison of a lacto-ovo-vegetarian diet (VD), Mediterranean diet supplemented with butyrate (MD-Bt) and Mediterranean diet without any supplement (MD), as control, within the same individual. Participants will act as their own controls in cross-over studies, so individual differences will be controlled for, making the error variance smaller and subsequently reducing the sample size required to find a significant effect due to increased statistical power. The study design follows the SPIRIT guidelines (see Figure 1 and supplementary file #1).
Eligibility criteria
Inclusion criteria include diagnosis of BS, age 18-65 years, willing to give informed consent, and willing to participate in a study where one of the proposed dietary profile is a vegetarian pattern.
Exclusion criteria include pregnancy or lactation, concomitant presence of serious illness or unstable condition (other immune-mediated or autoimmune diseases, including inflammatory bowel diseases); chronic viral infections; malignancies, recent myocardial infarction, chronic liver disease, current or recent (past 6 months) participation in weight loss treatment program or use of weight loss medication, and adoption of a vegetarian diet for the past 3 months. Antibiotic, prebiotic or probiotic use in the past 3 months.
Interventions and participant timeline
This clinical randomized study will use a cross-over design with three intervention periods and two wash-out. After a 2-week run-in period - which will be used to assess participants’ eligibility and to collect demographic details, signed informed consent and 3-day dietary records - the eligible participants (N=90) will be randomly assigned to follow a 3-month dietary profile with either: VD, MD or MD-Bt. The VD will contain inulin and resistant starch-rich foods, eggs and dairy, in addition to plant-based food, but will not contain meat, poultry or fish. The MD will contain all the food categories and will provide 2 portions per week of fish and 3 portions per week of fresh and processed meat (1 of which fresh or processed red meat). The MD-Bt will be similar to the MD but supplemented with 1.8 g/day of oral butyrate. The three different dietary patterns will be isocaloric and related to the participants’ nutritional requirements and will derive about 50-55% of energy from carbohydrates, <30% from fats, and 15-20% from proteins. Participants will prepare their meals or eat at restaurants. Alcoholic beverages will be limited to two per day for men and one per day for women. Interventions will be delivered by a dietitian through face-to-face, individual counseling sessions at the Careggi University Hospital. Participants will be provided with a detailed, 1-week menu plan with portions expressed in grams or milliliters as appropriate, as well as tips and information on the food groups that could be included and those that could not. The VD plan will also include recipes for preparing meals.
Study design is depicted in Figure 2. There will be six clinical evaluations of the study population: at baseline, before starting the nutritional interventions (T0), three months after the onset of the first nutritional intervention (T1), at the end of the first wash-out period, lasting 1.5 month, when individuals will be allowed to come back to their normal eating habits and at the onset of the second intervention (T2), at 7.5 months after the onset of the study and at the end of the second nutritional intervention (T3), at the beginning of the third nutritional intervention and at the end of the wash-out period, lasting again 1.5 month (T4), and at the end of the third nutritional intervention (T5).
During the baseline visit, participants will be instructed on the objectives and methods of the clinical trial and will be asked not to alter their physical exercise habits during the study. Anthropometric measurements, body composition, blood, urine and stool samples will be obtained from each participant at the beginning and at the end of each intervention phase. BS clinical manifestations and disease severity, especially for symptoms involving gastrointestinal system, will be assessed using a modified version of two questionnaires originally computed for Inflammatory Bowel Disease individuals: The Global Assessment of Improvement Scale (GAI) and The Symptom Severity Scale (SSS). Disease activity will be assessed by the validated Behçet Disease Current Activity Form (BDCAF), which consists of objective and subjective items, and considers the symptoms present over the 4 weeks prior to assessment. Each participant will have to complete a 3-day dietary record (two weekdays and a weekend day) before starting and a dietician will analyze all 3-day dietary records using a country-specific food-nutrient database.
Outcome measures
Primary outcomes
Primary outcomes will be assessed through validated questionnaires, in order to analyze BS progression. Used metric will be changes in means from the beginning to the end of each dietary intervention. They will include:
- Severity of gastrointestinal symptoms, assessed by the SSS modified form. The SSS is a multidimensional rating scale assessing overall symptoms' severity on a Visual Analogue Scale (VAS). An overall score will be calculated from six items: pain severity, pain frequency, abdominal bloating, bowel habit dissatisfaction, abdominal heaviness, and life interference. The modified SSS ranges from 0 to 600, with higher scores meaning more severe symptoms.
- Improvement of gastrointestinal-related BS symptoms, assessed by the GAI modified form. The GAI will assess BS improvement of symptoms using a 7-point scale, with higher scores meaning an improvement of the symptoms. The severity of abdominal pain, severity of abdominal distention, satisfaction with bowel habits, severity of headache, severity of exhaustion, severity of nausea, attention disorder, muscle/joint pain, and quality of life will be investigated in response to the following question: “Compared to the way you felt before you entered the study, have your symptoms over the past 7 days been: 1) “Substantially Worse”, 2) ”Moderately Worse, 3) ”Slightly Worse”, 4) ”No Change”, 5) ”Slightly Improved”, 6) ”Moderately Improved” or 7) “Substantially Improved”.
- Disease severity of BS, assessed by the BDCAF. The BDCAF will assess the presence of oral and genital ulceration, skin, joint and gastrointestinal involvement, presence of fatigue and headache with a 5-point scale according to the duration of symptoms, with 0 meaning no symptoms and 4 meaning symptoms for 4 weeks. The presence of eye, large vessel or central nervous system (CNS) involvement will be documented with "yes/no" answers. In addition, individuals will be asked to rate on a 7-point scale how active they felt. Similarly, clinicians will complete a 7-rating scale to assess their opinion of overall disease activity, with lower scores representing better outcomes.
Secondary outcomes:
Secondary outcomes will be measured in blood and stool samples. Used metric will be changes in means from the beginning to the end of each dietary intervention. Regarding stool samples, we will evaluate:
- Changes of GM composition from baseline, assessed by 16S rRNA gene-based next-generation sequencing on Illumina MiSeq platform. Total microbial DNA will be extracted from feces using the repeated bead-beating plus column method, as previously described [18]. The V3 and V4 hypervariable regions of the 16S rRNA gene will be sequenced following the Illumina protocol for 16S Metagenomic Sequencing Library Preparation.
- Fecal SCFA change from baseline, assessed by Gas Chromatography - Mass Spectrometry system (GC-MS). The metabolomic analysis of fecal waters will be performed after sample preparation involving solid phase microextraction (SPME), followed by GC-MS analysis to detect the volatile metabolites [4].
- Inflammatory profile change in feces from baseline, assessed by cytofluorimetric approach. In particular, superficial (CD3, CD4, CD8) and intracitoplasmic (transforming growth factor (TGF)-ß, interferon (IFN)-γ, interleukin (IL)-4, IL-9, IL-10, IL-17, IL-10, IL-22, FoxP3) cell marker analysis will be measured. The cells will be analysed by a BD FACScan Cytofluorimeter using the Diva software (BD Biosciences, San Jose, USA). The concentration of cytokines, chemokines and growth factors, including interleukins, IFN-γ, tumor necrosis factor (TNF)-α, growth-regulated oncogene-α (Gro-α), monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein-1ß (MIP-1ß), granulocyte colony stimulating factor (G-CSF) and granulocyte-macrophage colony stimulating factor (GM-CSF) will be determined in the fecal solutions according to the methodology of Munoz-González et al. [19].
In blood, we will evaluate:
- Inflammatory profile change from baseline, assessed by cytofluorimetric approach. It will be assessed with the Bio-Plex cytokine assay (Bio-Rad Laboratories Inc., Hercules, CA, USA), according to the manufacturer’s instructions. In particular, IL-1ra, IL-4, IL-6, IL-8, IL-10, IL-12, IL-17, MCP-1, MIP-1ß, vascular endothelial growth factor (VEGF), TNF-α, IFN-γ, IFN-γ-induced protein (IP)-10, will be measured.
- Lipid peroxidation markers' change from baseline, assessed by It will be estimated using the Thiobarbituric Acid Reactive Substances (TBARS) assay kit [20].
- Plasma total antioxidant capacity’s (TAC) change from baseline, assessed by fluorometry, using the oxygen radical absorbance capacity [21].
- Reactive oxygen species (ROS) change from baseline, assessed by flow cytometry. In particular, leukocyte subpopulations (lymphocyte, monocyte, and granulocyte) ROS will be measured [22].
In urine, we will evaluate:
- 1,4-dihydroxynonane mercapturic acid (DHN-MA) change from baseline, assessed by enzyme immunoassay using polyclonal antibodies [23]. This specific biomarker is the major urinary metabolite of 4- hydroxy-2-nonenal, a lipid peroxidation product.
Sample size calculation
Due to the lack of dietary intervention trials on BS and to the fact that gastrointestinal symptoms are quite similar in BS and in the inflammatory bowel disease, on the basis of a previously published trial [24], a sample size of 80 BS individuals was required to achieve power of 80% (beta) with alpha = 0.05, to detect a 50-points difference in the mean of SSS (the primary outcome), between VD and MD interventions. We will recruit 10 extra volunteers (for a total of N=90) as we assume that not all our individuals would be compliant with the treatments and in case of loss to follow-up. Losses will be included in the intention-to-treat but not in the per-protocol analyses.
Recruitment and randomization
Participants will be recruited from the Behçet Center and Lupus Clinic, Careggi University Hospital, Florence, Italy, or using advertisements on local media, newspapers, social media and websites. After approval and completion of the initial assessment, the participants will be formally included in the study and randomized with a 1:1:1 randomization to the three intervention arms through a web-based online randomization procedure. No adaptive randomization procedures will be performed. The random allocation sequence will be produced and managed by an independent staff member who is outside of the project, to code the treatments, and maintain the key to this code until data collection is completed.
Blinding
Although full blinding of both participants and dieticians to treatments in this study is not possible because of obvious differences between the intervention diets, several strategies will be employed to reduce the risk of bias. First of all, the treatment allocation for each individual will not be revealed until the individual has irrevocably been entered into the trial, to avoid selection bias. In addition, outcome measures in the present study cannot be easily influenced by the observer. Furthermore, trial personnel who will enrol participants, data collectors, outcome assessors and data analysts will be blinded to treatment allocation, and an employee outside of the research team will insert data into the computer in separate datasheets. On the other hand, making the trial open rather than blinded may improve recruitment. Unblinding will be permissible only when knowledge of the treatment will be absolutely essential for further management of the individual.
Data collection
Follow-up assessments and data collection will be undertaken at the Unit of Clinical Nutrition of the Careggi University Hospital, Florence, Italy, by trial personnel. All participants will be examined between 7.30 and 11.30 a.m. after a 12 hours-fasting period.
Compliance
Compliance with the interventions will be achieved using behavior change strategies including self-monitoring, and regular phone calls for dietary counseling. In particular, participants will receive at least one unannounced phone call during each intervention, in which participants will recall his or her 24-h diet period. Furthermore, participants will be provided with a detailed one-week menu plan for each dietary period with all foods expressed in weight and/or volume measures, and a hand-out containing details on their assigned diet, including food groups that can be included and ones that should be avoided. The vegetarian menu plan will include also recipes for preparing meals. Participants may discontinue the intervention or withdraw from the study for the following reasons: (1) at the request of the participant; (2) if the investigator considers that a participant's health will be compromised due to adverse events or concomitant illness that develop after entering the study. Participants prematurely discontinued from the study before the 3-month evaluation will have the baseline clinical and laboratory evaluations performed.
Anthropometric measurements and body composition
Weight and height will be measured using a stadiometer. BMI will be calculated as weight (kg) / height (m)2. Individuals will be classified as overweight if their BMI is more than 25 kg/m2 but less than 30 kg/m2, and obese if their BMI is 30 kg/m2 or more. Body composition will be determined with a bioelectrical impedance analyzer device (TANITA, model BC 420 MA) at the beginning and at the end of each intervention phase.
Blood samples
Blood samples will be collected at each clinical evaluation. Blood samples will be centrifuged at 3,000 rpm for 15 minutes to yield serum, aliquoted, and then stored at -80°C until analyses.
Urine and stool samples
Urine and fecal samples (four or five scoops totaling 4 g) will be collected in sterile containers with no solution before and after each intervention phase - for a total of six samples for each participant - and immediately frozen at −20 °C, before being transferred to −80 °C until analysis. Urine and stool sample collection kits, including containers and instructions, will be provided for the participants. Participants will be invited to place the stool samples in a supplied sterile container with no solution, placed in a plastic bag and put into a household refrigerator immediately. Within 24 h, urine and stool samples will have to be delivered to the laboratory using a styrofoam box containing ice packs.
Storage of biological specimens
The storage of biological specimens will be performed under appropriate conditions, according to standard methods. Blood samples will be aliquoted and stored at −80°C for 4 years before being used or destroyed. Stored samples will be used exclusively for research purposes upon consent of the donor. Sample destruction will be appropriately documented.
Data management
Data will be collected on an electronic database. Identifiable data or other documents will not be recorded in the database and participants will be identified by a unique trial ID only. Hard copies of data sheets linking the participant identification number to the person's contact details will be kept securely in a locked filing cabinet in a locked office, accessible only to key research team members. Participant files and other source data (including copies of protocols, questionnaires, original reports of test results, correspondence, records of informed consent, and other documents pertaining to the conduct of the study) will be kept for the maximum period of time permitted by the institution. All data of the project will be stored in the DASH-IN infrastructure, which is developed by ENPADASI. Thereby we will adhere to the FAIR principles (Findable, Accessible, Interoperable and Reusable). The data will be made open access upon publication. Within Europe the ELIXIR infrastructure is coordinating data stewardship and management activities in the life sciences.
Multiple strategies will be employed to improve data quality during data collection, including an accurate recruitment, a structured and time-limited protocol, the inclusion of a run-in period, the limitation of the burden and inconvenience of data collection on the participants, the development of a trusting and collaborative relationship between research units and participants and double data entry.
Statistical analysis
The results will be expressed as mean ± SD, median and range, or geometric mean with 95% confidence intervals (CIs) as appropriate. Categorical variables will be presented in terms of frequencies and percentages. All data will be treated as paired samples from a crossover study. The three interventions will be analyzed combining the results obtained in the three phases of the three groups. Outcomes will be analyzed within each group using paired comparison Student’s t-tests to test whether the changes will be statistically significant. The absolute change (mean baseline value subtracted from mean value after each intervention) will be estimated with independent t sample tests.
For the GM analysis, raw sequences will be processed using a pipeline that combines PANDAseq [25] and QIIME [26]. The UCLUST software [27] will be used to bin high-quality reads into operational taxonomic units (OTUs) at 0.97 similarity threshold through an open-reference strategy. Taxonomy will be assigned through the RDP classifier, using the Greengenes database as a reference. Alpha-diversity and rarefaction curves will be computed using different metrics, including the Faith's phylogenetic diversity, the number of observed OTUs and the Shannon index. Beta-diversity will be estimated by weighted and unweighted UniFrac distances, which will be used as input for Principal Coordinates Analysis (PCoA). All statistical analysis will be performed using R and the packages vegan, stats and made4. The significance of data separation in the PCoA will be tested using a permutation test with pseudo-F ratios (function Adonis of vegan). Wilcoxon test for paired data will be used to assess significant differences in alpha diversity and taxon relative abundance between groups, while Kruskal–Wallis test will be used for multiple comparisons. P-values will be corrected for false discovery rate (FDR, Benjamini-Hochberg).
Permutational analysis of variance (PERMANOVA) adjusted for the order of treatment will be applied with two objectives: to assess the effect of the dietary interventions on PCoA scores of beta diversity metrics and to assess the effect of the dietary interventions and of several metadata (SCFA’ levels, nutrients’ intake, clinical and biochemical parameters) on the whole microbial composition. In the last chance, the analysis will be performed at each taxonomic level (Genus, Family, Order, Class and Phylum), with a separated model for each metadata. Each analysis will be performed using a model including both the main effects and the interaction of the three dietary interventions. Because these tests assume normal data distribution, non-distributed data will be transformed into logs, and further analyses will be performed with the processed data. However, to facilitate interpretation, the log data will be again converted to the original scale (antilog) and presented as geometric means with 95% CIs. Spearman correlation coefficients between the changes in microbial composition, SCFA’ levels, nutrients’ intake and in clinical and biochemical parameters will be computed and statistically tested for each dietary intervention.
Before starting the data analysis, the level, pattern and likely causes of the missingness in the baseline variables and outcomes will be investigated by forming appropriate tables. This information will be used to determine whether the level and type of missing data has the potential to introduce bias into the analysis results or substantially reduce the precision of estimates for the proposed statistical methods. Sensitivity analyses will be undertaken, based on the assumption that missing outcomes are the worst possible, or the best possible, in different randomization groups. If these show that conclusions may differ based on missing values, then supplementary multiple imputation for missing values will be undertaken. These analyses will account for results of any losses to follow-up insofar as they pertain to differences in measured variables (i.e. under the assumption of missing at random). Finally, subgroup analyses will be performed to analyze possible differences in the changes according to some characteristics of the study population, such as age, sex, categories of BMI and physical activity. A p value < 0.05 will be considered statistically significant. Statistical analyses will be performed using SPSS software for Macintosh (SPSS Inc., Chicago, IL, US).
Monitoring
Given the limited objectives and its short-term nature, this trial will be monitored on a regular basis by the protocol team and the local Institutional Review Board, without the use of a formal data monitoring committee. Each month, the protocol team will provide the local Institutional Review Board with a monitoring report, including a review of activities, progress, difficulties and issues of concern. No ad interim analysis will be performed. Data access will be restricted to trained staff with unique password-protected accounts. Adverse events such as unfavourable and unintended signs, abnormal laboratory findings, symptoms, or diseases temporally associated with the intervention diet will be collected from the time of randomisation until the final 12-month follow-up visit for each participant, whether or not considered related to the intervention study. All adverse events will be followed up until they are resolved.