Study Protocol for the Effects of Formula Diet with Dapagliflozin on Metabolic Improvement and Body Composition in Type 2 Diabetes Mellitus

The study design is a prospective, multicenter, double-blinded, randomized, controlled, investigator-initiated clinical trial. Subjects are randomized to two groups: high protein FD (group A) and high fat FD (group B). The two FD have the same number of calories (182 kcal) and are started simultaneously with dapagliflozin. In group A, patients take 5 mg of dapagliflozin orally once daily and replace one meal with a high protein FD while taking two normal meals a day. In group B, patients take dapagliflozin 5 mg orally once daily and replace one meal with a high fat FD while taking two normal meals a day. The observation period for both groups is 24 weeks. Patients are recruited from seven facilities participating in this study. Eligible patients who provide written informed consent are randomized to group A or group B at a ratio of 1: 1 by a computer program located at the registration center, using a minimization method balancing age (≤ 65 or > 65 years), HbA1c level (≤ 8.0 or > 8.0%), and waist circumference (men, ≤ 85 or > 85 cm; women, ≤ 80 or > 80 cm) at the time of screening (Fig. 1).

The primary endpoint of this study is the change in HbA1c. Shirai et al. [11] reported a change in HbA1c of − 0.6 ± 1.1% using a high protein FD. According to a domestic phase III clinical trial of dapagliflozin, the change in HbA1c after 24-week treatment with dapagliflozin is − 0.4% ± 0.7%. Although there is no report on high fat FD, the protein/fat/carbohydrate (P/F/C) ratio per day in group B is not so different from conventional Japanese food. We thus estimate that dapagliflozin treatment will result in the same level of change in group B. Assuming that an additive effect is obtained in group A using a high protein FD and dapagliflozin, the change in HbA1c at 24 week is estimated to be − 1.0%. In group B using a high fat FD and dapagliflozin, the change in HbA1c will be similar to that of the domestic phase III clinical trial of dapagliflozin and is estimated to be − 0.6%. The standard deviation of the change in HbA1c is presumed to be 0.8%, which is the range in general clinical studies. At a significance level (α) of 0.05 and a detection power [100(1 − β)] of 80%, the number of cases per group necessary to detect a significant difference in change in HbA1c between the two groups is calculated to be 64 cases. Assuming 5% deviant and omission samples, the target number of cases for each group is 68 cases, with a total of 138 participants in this study.

Sunny Health Co., Ltd. is the manufacturer and distributor of the FD. The number of calories of a normal meal (conventional Japanese food) is calculated by 35 kcal/kg × standard body weight (kg). We assume that the standard body weight is equivalent to a body mass index (BMI) of 22 kg/m². For example, if the patient’s standard body weight is 60 kg, a 35 kcal/kg meal would be 2100 kcal per day, which is equivalent to 700 kcal per meal. Therefore, if the subject takes FD (182 kcal) once a day and a normal meal of 700 kcal twice a day, the total calorie will be 1582 kcal per day, which is equivalent to 26.4 kcal/kg.

The protein components of FD consist of egg white, casein, and soybean protein. The P/F/C ratio is approximately 42:12:46 for the high protein FD (group A) and 2:55:43 for the high fat FD (group B). When the subject’s standard body weight is 60 kg, the overall P/F/C balance is 21:24:56 for group A and 16/28/56 for group B.

Since the height and standard body weight of each subject differ, the P/F/C ratio also differs among subjects, but the errors are acceptable. For example, when the height of the subject is 140 cm vs. 190 cm, the P/F/C balance in the high protein FD group is 22:23:55 vs. 20:24:56, and that in the high fat FD is 16:29:55 vs. 17:27:56.

The FD is reconstituted in a 600-ml dedicated shaker. A package of the FD together and 350–400 ml of water are shaken well before consumption. On weeks 0 and 12, subjects receive instructions on how to use the formula diet and nutritional guidance about calorie intake from a nutritionist at each facility. In the absence of a nutritionist, the instructions are given by nurses.

When a low carbohydrate diet is given to a patient using SGLT2 inhibitors, the onset of diabetic ketoacidosis (DKA) is a concern. In this study, we use FD which is a low-calorie diet itself, but the use of FD is limited to one pack per day. For the remaining two meals, nutritional guidance targeting 35 kcal/kg is given, and considering on a per day basis, both groups will take about 25 kcal/kg. Also, carbohydrates are taken at about 56% according to P/F/C ratio. A low carbohydrate diet of less than 50 g per day is considered to be a ketogenic diet [18]. The daily calorie intake of participants in this study is determined by the height of the individual, but even a person with a height of 140 cm has about 160 g of carbohydrates intake per day. Therefore, we consider that diet therapy in this study is not an intense carbohydrate-restricted diet even considering urine glucose excreted by treatment with dapagliflozin. Also, blood and urinary ketone bodies are monitored during the observation period.

This study is conducted at Toho University Sakura Medical Center and six other hospitals or clinics. Participating facilities are listed in https://​upload.​umin.​ac.​jp/​cgi-open-bin/​ctr/​ctr_​view.​cgi?​recptno=​R000028290. This study was registered in the University Hospital Medical Information Network (UMIN) Clinical Trials Registry (UMIN000024580) on March 1, 2017. The trial is ongoing and is scheduled to be completed in June 2019.

This study is conducted with a double-blind design, and blinding is maintained throughout the study for all research stakeholders except the allocation officer through the following procedures. The investigator fills in the necessary information of a potential subject in the case registration card using subject identification number and transmits by facsimile to the research support organization (Ibec Corporation, Osaka, Japan). The allocation officer of the research support organization confirms eligibility of the potential subject. When eligibility is confirmed, the allocation officer randomizes the subject according to an allocation table consisting of the serial number, the test food group identification number in random order, and the test food allocation code. The allocation officer replies by facsimile to the investigators regarding the results of eligibility and randomization. The allocation officer ensures that the appearance of the test food, the packaging, and the indications on the label of the two FD are indistinguishable. The allocation officer keeps the allocation table sealed until the key is opened, and does not disclose it to other related persons involved in the study except during emergency. When unused test food is collected before key opening, the food is collected in a sealed state to maintain blindness.

After registration and randomization, dapagliflozin 5 mg is taken orally once daily for 24 weeks. In principle, a daily dose of 5 mg is taken, but if HbA1c exceeds 8.5%, the dose may be increased to 10 mg at the investigator’s discretion. At the same time, the subjects take a high protein or high fat FD to replace one of the three meals a day.

The schedule of study visits and data collection is summarized in Table 1.

The investigators confirm the status of compliance to medication and test food intake at each visit and instruct the subjects to bring empty packages of the test food and unused test food at the next visit.

In principle, treatments for diabetes, hypertension, dyslipidemia, and hyperuricemia are not changed from the time of obtaining consent to the completion of study, to avoid effects on the efficacy and safety assessments of this study. However, changes are permitted when the glucose level improves markedly owing to sufficient weight loss or when the onset of hypoglycemic attacks is a concern. If it is judged that adding a new antidiabetic drug is necessary because of insufficient therapeutic effect, the trial will be discontinued at that time. If treatment is changed, the contents of change and the reason and timing are described in the case report form.

The primary endpoint is the change in HbA1c from the start of treatment to 24 weeks after initiation of treatment, and is compared between the two groups. Secondary endpoints are the rates of change in (1) body weight and BMI; (2) blood pressure and heart rate; (3) waist circumference; (4) fasting blood glucose level and insulin level for the calculation of homeostatic model assessment (HOMA)-β and HOMA-R; (5) serum lipid levels (total cholesterol, LDL-cholesterol calculated using the Friedewald equation, fasting triglyceride, HDL-cholesterol); (6) ketone fraction (venous blood); (7) lipid-related parameters [pre-heparin lipoprotein lipase (LPL) protein, lipoprotein fraction (polyacrylamide gel electrophoresis), remnant-like lipoprotein cholesterol, apoprotein (Apo A1, A2, B, C2, C3, E)]; (8) high sensitivity C-reactive protein (hs-CRP); (9) urinary albumin quantification; (10) ankle brachial index (ABI) and cardio-ankle vascular index (CAVI) as a marker of arterial stiffness (only when measurement is possible), (11) body composition (only when measurement is possible). Insulin, ketone fraction, lipid-related parameters, and hs-CRP will be measured at a central laboratory (LSI Medience Corporation, Tokyo, Japan). ABI and CAVI are measured using the VaSera series (FUKUDA DENSHI, Tokyo, Japan). Body composition is evaluated by the InBody series (Biospace, Tokyo, Japan) that performs multifrequency bioelectrical impedance analysis. Serious adverse events (AEs) will be documented and reported per regulatory requirements.

A novel arterial stiffness diagnostic parameter called CAVI has been developed in Japan, which essentially reflects the stiffness of the aorta, femoral artery, and tibial artery [19]. This stiffness parameter has been reported to be less affected by blood pressure at the time of measurement and adequately reproducible for clinical use [20, 21]. CAVI is also known to be associated with traditional parameters of vascular structure and function, such as intima–media thickness (IMT), pulse wave velocity (PWV), and central augmentation index [22]. Furthermore, no special technique is required for the measurement of CAVI. CAVI has been associated with a number of risk factors for cardiovascular disease (CVD) and the severity of CVD [23‐27]. An increase in number of risk factors for metabolic syndrome also correlates positively with increased CAVI, probably as a result of insulin resistance [28].

Body fat is measured by bioelectrical impedance analysis (BIA) using Inbody. The BIA method is noninvasive and easy to measure and is more convenient and effective when measuring multiple times. The BIA method measures the resistance (impedance) of a body by applying a weak and noninvasive current to the body. Since the muscle contains water that readily conducts electricity, the impedance decreases as the amount of muscle increases, and conversely the impedance increases as fat increases. Impedance in the body is measured and the impedance index is calculated by an established formula [29]. Body water content is calculated from this impedance index, and lean body mass is determined on the basis of the fact that approximately 73% of the fat-free mass is body water. Further, the body fat mass is obtained by subtracting the fat-free mass from the body weight.

The investigators will maintain the records of individual patients as source data, which include a log of informed consent forms, medical history, laboratory data, and other records, as appropriate. All entries in the case report forms (CRFs) should be backed up by the relevant source data. CRFs must be completed in a timely manner.

All data will be collected by the data management center. There will be no direct communication between investigators and data managers. Clinical data entry, coding, data management and reporting will be performed by the research support organization, with reading and visual check, and a data clarification form (DCF) is issued if there is inconsistency.

Monitoring and auditing of this research will be conducted by Clinical Research Office Corporation (Kanagawa, Japan). Monitors will ensure that the investigational team complies with the study protocol, that data and AEs are accurately recorded in the CRFs, that severe AEs are forwarded to the trial coordinator and the investigational drug provider, and that those meeting reporting criteria are forwarded to the institutional review board (IRB).

When AEs occur, the investigators immediately take appropriate measures and describe the AEs in the medical record and CRFs without inconsistency. For investigation of adverse events, the investigators conduct follow-up until the subjects recover to the state before initiation of treatment and confirm the outcomes. If the follow-up cannot be performed for any reason, this should be described in the CRFs and the investigation will be terminated.

In the case that identification of the test food is necessary to avert an emergency, the investigator (or the subinvestigator if the investigator is absent) at the study site requests unblinding of the allocation code by the test food allocation manager, through the principal investigator, and receives the result of disclosure for that case. Disclosure of information after emergency code opening is determined elsewhere.

Audit personnel are independent of those that perform monitoring and quality control. The person in charge of auditing conducts audits for quality assurance in accordance with the audit procedures prepared prior to implementation of audit.

Analyses of primary and secondary endpoints are performed using full analysis set (FAS) and per protocol set (PPS). As a rule, analyses are carried out under the condition of a 95% confidence interval (CI). A medical statistician prepares a statistical analysis plan and specifies details of the statistical methods including data handling. The statistical analysis plan is prepared before data fixation.

Subject to patient consent regarding use of the analytical data, the statistician checks the data of individual patients for any deviation from the protocol and the degree of deviation, as well as any dropout from the study and the timing of dropout. The statistician then defines the following three sets to be analyzed. The FAS contains cases that satisfy the following: (i) diagnosed as type 2 diabetes, (ii) randomized and received test treatment at least once, (iii) has data on efficacy after starting the study. The PPS contains cases showing no violations of selection criteria, exclusion criteria violation, and concomitant drug, with compliance not less than 75% and not more than 120%. The safety analysis set contains cases that are enrolled in the study and have started and received a part or all the test treatments.

Summary statistics of the background data are calculated for the test groups. Nominal variables are expressed as frequency or percentage of each category. Summary statistics (number of cases, mean, standard deviation, minimum, median, and maximum) are calculated for continuous variables. Patient characteristics are compared using a χ² test or Fisher’s exact test for categorical variables, and Student’s t test or Wilcoxon rank sum test for continuous variables. In the analyses of primary and secondary outcomes, summary statistics (number of cases, mean, standard deviation, minimum, median, and maximum) are calculated for the measured values, the amount of change, and the rate of change for each evaluation item. The amount of change and the rate of change are compared using one-sample t test for intragroup comparison and two-sample t test for intergroup comparison. The correlation between HbA1c and each variable is then evaluated using Pearson’s moment correlation coefficient or Spearman’s rank correlation coefficient.