This was a Phase 3, prospective, randomized, double-blind, placebo-controlled, multicenter clinical trial to evaluate the efficacy, safety, and tolerability of D-tagatose (as compared to placebo). Prior to the start of treatment, subjects participated in an 8-week run-in period during which diabetes education was provided and diet and exercise treatment stabilized. Efficacy analyses were conducted on data from 4 and 6 months.
Screening Visit
Based on the inclusion and exclusion criteria, eligible subjects underwent a basic physical examination, pregnancy test for females, and hematology, clinical chemistry, and urinalysis tests. Basic physical examination included physical measurements, general examination by observation, palpation, percussion, auscultation, blood pressure measure, and heart rate check. The clinical laboratory tests included: 1) hematology (hematocrit, hemoglobin, MCH, MCHC, MCV, total white blood cells, platelets, and differential); 2) clinical chemistry (sodium, chloride, potassium, CO2, BUN, uric acid, albumin, creatinine clearance, SGOT, SGPT, bilirubin (total and direct), phosphorus, calcium, alkaline phosphatase, total protein, and glucose (fasting); 3) HbA1c; 4) serum lipid profile including total cholesterol, HDL, LDL, and triglycerides; 5) urinalysis (appearance, volume, specific gravity, pH, glucose, protein, and microscopic evaluation of urinary sediment).
Run-in Period
Subjects deemed potentially eligible on the basis of their medical records were to continue on their previously established diet and exercise treatment program under physician's recommendation for 8 weeks. Patients were to record their food intake and exercise in nutrition diaries.
Treatment Period
The treatment period consisted of 12 monthly visits, the first of which was used to gather the baseline data for the efficacy and safety parameters and also included the first distribution of test and placebo treatments. Subjects who received study drug for more than 6 months were evaluated for safety and efficacy the first 6 months and then for safety only thereafter.
Baseline and Distribution of Test and Placebo Treatments
Approximately 550 enrolled subjects were to be randomized into the study. Subjects were randomized to one of the two treatments in a 1:1 ratio. Each eligible subject was randomized to a group assignment on the basis of a randomly generated number. Randomization was stratified by the screening HbA1c values (\<7.5% and ≥7.5%) to ensure an even distribution of subjects between the two treatments. Randomization was to take place through an Integrated Web-based Randomization System (IWRS) which sites accessed in order to make randomized treatment assignments.
After the 8 weeks run-in period, subjects returned (Visit 2) to the study sites and underwent medical history review, and the same examination and clinical laboratory tests as those performed during the screening visit. In addition, subjects received their randomized study treatment and detailed instructions about its use.
There were two treatment groups in this trial: 1. Drug (D-tagatose), and 2. Placebo (Splenda). The dose of tagatose was 15 g in solution in a final volume of approximately 4 ounces of water (\~250 mL) taken 3 times daily (tid); the dose of placebo (Splenda) was to be 1.5 g.
Treatments were dispensed by an unblinded staff member who had no other role in the trial; all other clinical center staff were blinded to treatment. If the patient was outside the +/- 7 day window, the investigator/site staff would document the information in the subject chart and provided the Sponsor with a note to file.
Procedures Visits
Every other month after Visit 2 subjects returned to the study site and underwent the same comprehensive testing that they underwent at Visit 2. Drug and dietary compliance was to be verified by sachet counts at each visit and by nutritional diaries.
DATA QUALITY ASSURANCE
A study initiation meeting was held at each investigative site before enrollment of the first subject to review study design, study procedures, data collection, and investigator responsibilities. During interim field monitoring visits, the monitor verified CRF entries against available source documents and ensured that appropriate data were collected and documented. Independent GCP audits of selected study sites were performed. Prior to data entry, CRFs were checked for completeness and consistency by a monitor and verified against source documentation that was maintained at each study site. Data from CRF were reviewed and queries to the study sites were generated by the clinical and data management staff of the CRO. Discrepancies were reconciled with the study site and the respective laboratory.
Prior to data lock a final audit report and memo was generated summarizing the audit findings and given to the Data Manager Data analyses were performed primarily with Statistical Analysis System (SAS®) software, version 9.2. SAS analysis datasets were crosschecked against the locked database and/or CRF. Data summary tables were crosschecked against SAS analysis datasets. All identified inconsistencies were reported to the lead statistician and resolved prior to production of this report. This clinical study report was checked for accuracy and consistency before sign-off.
Statistical and Analytical Plans Amarex LLC, Germantown, MD prepared a statistical analysis plan (SAP) for the efficacy and safety data before database lock.
Analysis Populations
Intent-to-Treat Population
The intent-to-treat (ITT) population was defined as all randomized subjects who received at least one dose of double-blind study medication and had a least one post-treatment visit evaluating efficacy. The ITT population was used for the main efficacy evaluation.
Per Protocol Population
The per protocol (PP) population was defined as all subjects in the ITT population who had 80% compliance with medication for 75% of the dosing time points and who had no major protocol violations or eligibility violations.
Safety Population
The safety population consisted of all randomized subjects who received at least one dose of double-blind study medication and had at least one post treatment visit evaluating safety. This population was used for all of the safety analyses.
Handling of Duplicate, Missing and Outlier Data
For data that were identified by the Sponsor's Medical Monitor as clinical outliers, two separate analyses were conducted, one including the outlier and one from which the outlier was excluded. The impact of all possible combinations of the outliers was to be treated in the context of the parameter. For the ITT population, missing data (including missing values at intermediate visits) were imputed from scheduled visits using the last-observation-carried-forward (LOCF) method. A blind review of the data was conducted prior to statistical analysis to determine data accuracy. Safety data were compiled on all subjects without exclusion, and in the case of AEs, the relationship to study product was made. The determination that a subject was to be excluded from an analysis population was jointly made by the sponsor and the investigator in the blind review meeting.
Covariates, Stratification Factors, and Subgroups
To improve the precision of the analysis, covariate analyses were conducted. The respective baseline values were used as covariates in the analysis models for all the continuous variables; along with the pre-specified stratification factor, screening HbA1c values (\<7.5% and ≥7.5%). The adjusted least squared (LS) means at each time point were reported. Other clinically meaningful baseline covariates and stratification factors were examined; if statistically significant, these factors were included in the analysis models.
Interim Analyses
Two blinded interim analyses (IAs) were planned, one when approximately 40% of subjects had been randomized and completed through month 6, and the other when over 60% of the subjects had been randomized and completed through month 6. The ITT population was used for the analyses conducted in each IA. The IA was directed and conducted by an independent statistician, who was not involved in the trial conduct other than this IA. The main purpose of the IAs was to assess the assumption made concerning the standard deviation in the sample size calculation for the primary end-point (mean change in HbA1c). Two such IAs were planned because the trial started earlier in the USA than in India, so the first IA was more likely to reflect the results from the USA population, while the second IA was more likely to reflect the results from the India population. For each IA, the independent statistician used the blinded data, the observed pooled standard deviation, and the other original assumptions noted in the sample size determination (e.g. Power, Type I error rate) to recalculate the sample size. Based on the documented sample size re-calculation and the observed dropout rate, the intent was that the independent statistician could recommend a new, larger sample size to the sponsor; however, in neither IA was this necessary.
Statistical Methods Subject Characteristics
The subject population enrolled in this trial was characterized according to
* the disposition of subjects within the trial;
* the demographics, baseline clinical laboratory assessments, and previous and concomitant medical history;
* concomitant medications; and
* treatment duration and compliance characteristics.
Demographic and baseline characteristics were summarized and also provided as data listings. Continuous data variables were summarized according to the number of observations, mean, standard deviation, median, and minimum and maximum values, while categorical data variables were summarized according to frequency counts and percentages.
Treatment duration and compliance characteristics were summarized by treatment group and analysis population with the duration of dosing (days) being calculated as date of last dose minus date of first dose plus 1. Percent compliance was calculated as \[(total units of drug returned minus the total units of drug dispensed) divided by the total dispensed\] times 100%.
Efficacy Analyses
Statistical tests for the efficacy analyses were to be two-sided, with no p-value adjustment, and were to focus on the responses of the ITT population with the PP population being used in supportive analyses. To assess the primary end-point of change in HbA1c level from baseline, the general linear model (analysis of covariance) adjusted for baseline and important stratification factors, e.g., baseline HbA1c and site, was used. In this analysis, in order to protect the statistical degrees of freedom, sites were pooled according to location and other factors. The closed test procedure was to be used for the secondary end-points in order protect the Type I Error rate of α = 0.05. For this procedure, the secondary end-points were to be analyzed in the following order: 1. Body Mass Index 2. Triglycerides 3. LDL 4. Total cholesterol 5. Fasting blood glucose 6. Proportion of subjects achieving HbA1c \<7% and/or ≤6.5% 7. Insulin 8. HDL
