Study design and subjects
We undertook a randomized, double-blind, placebo-controlled, parallel arm study with the first subject enrolled on 30 March 2015 and the last subject completed the study on 19 February 2016. The study was conducted in the Diabetes and Endocrine Research Centre of the Chinese University of Hong Kong (CUHK) at the Prince of Wales Hospital, Hong Kong Special Administrative Region. Subjects were identified from non-specialist general medical or family medicine clinics at the hospital. We recruited Chinese subjects aged between 18 and 70 years inclusive, fulfilling at least two of the following three criteria: 1) fasting plasma glucose 5.6–5.9 mmol/L (IFG) and/or 2-h plasma glucose 7.8–11.0 mmol/L (IGT) during a standard 75 g oral glucose tolerance test (OGTT); 2) glycated haemoglobin (HbA1c) 5.7–6.4% (39–46 mmol/mol); and 3) at least one of the following risk factors of a) history of gestational diabetes, b) history of diabetes in first degree relatives, and c) two or more of metabolic syndrome components of triglyceride ≥1.7 mmol/L, high density-lipoprotein (HDL) cholesterol < 1.3 mmol/L in women or < 1.1 mmol/L in men, waist circumference ≥ 80 cm in women or ≥ 90 cm in men, or blood pressure (BP) ≥130/80 mmHg. Exclusion criteria included current use of dietary supplements known to affect glucose or galactose metabolism, use of anti-diabetic medications in the previous 6 weeks, cardiovascular disease in the recent 12 months, renal impairment with estimated glomerular filtration rate < 60 mL/min/1.73m
2, history of eating disorder, and known lactose or galactose intolerance. The study was registered at
www.clinicaltrials.gov, reference number NCT02358668.
Intervention
Subjects meeting eligibility criteria were randomly assigned to receive BTI320 4 g (n = 24), BTI320 8 g (n = 24) or matching placebo (n = 12) orally three times daily, 10 min before each main meal for 16 weeks. Each 4-g tablet of BTI320, administered as a chewable tablet, contained 2.0 g of the key ingredient mannan polysaccharides. Other ingredients included food grade sorbitol, magnesium stearate, malic acid, natural flavors and colors. Subjects were instructed to maintain their usual dietary pattern and physical exercise levels. Subjects were reviewed every 4 weeks for assessment of adverse events and drug compliance, the latter was established by counting the returned tablets.
Clinical measurements
Serum fructosamine was measured at baseline and 4-weekly interval until completion of treatment at 16 weeks, HbA1c at baseline and 16 weeks, and OGTT at screening and 30-day post-treatment visit. Meal tolerance test (MTT) using a standardized meal of 500 kcal was conducted at baseline, 4 weeks and 16 weeks measuring plasma glucose, insulin, C-peptide and glucagon-like peptide (GLP)-1 at 0, 15, 30, 60, 90 and 120 min. The standard meal consisted of two pieces of pineapple shortcakes and one carton of soymilk with nutritional breakdown as follows: carbohydrates 75.7 g (57.1% of total energy intake), fat 21.2 g (35.9% of total energy intake) and protein 9.3 g (7.0% of total energy intake). Seventy-two-hour CGM using the Medtronic iPro®2 CGM and Enlite sensor was performed at baseline, 4 weeks and 16 weeks. Other metabolic parameters (body weight, waist circumference, BP, lipid [total cholesterol, HDL-cholesterol, triglyceride, low density-lipoprotein cholesterol], high-sensitivity C-reactive protein [hs-CRP]) and safety parameters (renal function, liver function and complete blood count) were measured at regular intervals.
Fructosamine was measured using colorimetric test by reaction with nitroblue tetrazolium. The measuring range of the fructosamine assay was 14–1000 μmol/L, intra-assay coefficient of variations (CVs) were 0.8% and 0.5% at concentration of 275 μmol/L and 515 μmol/L, respectively, and inter-assay CVs were 1.5% and 1.2% at concentrations of 262 μmol/L and 489 μmol/L, respectively. Glycated haemoglobin was measured using immunoassay traceable to the National Glycohaemoglobin Standardisation Program and the International Federation of Clinical Chemistry standards. The measuring range of HbA1c assay was 0.3–3.4 g/dL, inter-assay CVs were 1.2% and 0.7% at concentrations of 5.3% Hb and 9.6% Hb, respectively, and inter-assay CVs were 2.2% and 1.9% at concentrations of 5.0% and 10.4% Hb, respectively. Insulin was measured by immunoassay which had a measuring range of 2–300 mIU/L with intra-assay CVs of 3.6% and 2.9% at concentrations of 11.7 mIU/L and 51.2 mIU/L, respectively, and inter-assay CVs of 6.7% and 5.3% at concentrations of 11.2 mIU/L and 47.4 mIU/L, respectively. C-peptide was measured using immunoassay which had measuring range of 0.1–20 μg/L with intra-assay CV of 2.8% and 1.7% at concentrations of 0.7 μg/L and 6.2 μg/L, respectively, and inter-assay CVs of 3.5% and 6.3% at concentrations of 0.8 μg/L and 6.3 μg/L, respectively. Glucagon-like peptide 1 was measured by enzyme-linked immunosorbent assay (Immuno-Biological Laboratories Co. Ltd., Japan). The measuring range of GLP-1 was 1.25–80 pmol/L, intra-assay CVs were 9.8% and 2.2% at concentrations of 5.0 pmol/L and 7.8 pmol/L, respectively, and inter-assay CVs were 10.3% and 5.7% at concentrations of 6.1 pmol/L and 11.0 pmol/L, respectively. Glucose, total cholesterol, HDL-cholesterol and triglyceride were measured using the enzymatic colorimetric method. Insulin and C-peptide were analysed by the Siemens IMMULITE® 2000 XPi Immunoassay System, HbA1c was measured on the Roche Cobas Integra 800 System (Roche Diagnostic GmbH, Mannheim, Germany), GLP-1 was measured manually, and the rest of the assays were measured on the Roche Cobas c8000 Analytical System (Roche Diagnostic GmbH, Mannheim, Germany). All laboratory tests were performed in the Department of Chemical Pathology, the CUHK, the Prince of Wales Hospital, which was accredited by the National Association of Testing Authorities, Australia and the Royal College of Pathologists of Australasia for medical testing.
All subjects completed Food Frequency Questionnaire, Hill and Blundell questionnaire on appetite, International Physical Activity Questionnaire, and World Health Organisation Quality of Life questionnaire at baseline, 4 weeks and 16 weeks.
Efficacy endpoints
The primary endpoint was change in serum fructosamine in subjects treated with low dose and high dose BTI320 compared with placebo from baseline to 4 weeks. The main secondary endpoints were changes in calculated indices of glycaemic variability (mean post-prandial incremental AUC [AUCpp] at 1 h, 2 h and 3 h, mean post-meal maximum glucose [MPMG], AUC-180, mean amplitude of glucose excursion [MAGE], standard deviation [SD], and percent CV) based on CGM data in subjects treated with low dose and high dose BTI320 compared with placebo during the study. The AUCpp is the area above pre-prandial glucose starting from the beginning of each main meal to 1 h, 2 h and 3 h after the meal, obtained using the trapezoidal rule. The MPMG is the mean maximal glucose value within 3 h after each main meal. The AUC-180 is the AUC for glucose level above 180 mg/dL (10 mmol/L). The MAGE is the mean difference in glucose values between consecutive peaks and nadirs, only considering changes above and below mean glucose of more than 1 SD [
19]. The percent CV is SD divided by mean glucose values. Other secondary endpoints included changes in HbA1c, 2-h AUC of plasma glucose, insulin, C-peptide and GLP-1 post-MTT, body weight, BPs, lipids, hs-CRP, as well as changes in self-reported dietary intake and satiety from baseline to end of treatment in subjects treated with low dose and high dose BTI320 compared with placebo.
Statistical analysis
In our estimation of sample size, we assumed a mean serum fructosamine level of 273 μmol/L with SD of 22.5 μmol/L in the placebo arm, and a change of 10% in fructosamine level would be detected using a two-sided 5% level test with 80% power if there were 11 subjects per arm.
Efficacy analyses were performed in the intention-to-treat population which consisted of all randomized subjects who have received at least one dose of the assigned treatment. A
per protocol analysis was also performed in subjects who have taken at least 70% of the treatment. Analysis of covariance (ANCOVA) was used to measure the changes in serum fructosamine, and changes in other glycaemic and metabolic indices from baseline to week 4 and week 16 between intervention arms, adjusted for age, gender and baseline measurements. The effects of low or high dose BTI320 compared with placebo on CGM glycaemic variability indices were further explored using random effect models with repeated measurements adjusted for intra-individual between-meal and between meal-day variability, age and gender. Linear mixed effect is a common statistical method to address repeated measurements [
20,
21]. Post-hoc subgroup analysis was conducted on significant CGM glycaemic variability indices by dividing the population into 1) Low and high body mass index (BMI) stratified by the population BMI median; 2) Younger and older age groups by population age median; and 3) With IFG and/or IGT at baseline and without IGF and IGT at baseline. Analysis was performed using Statistical Analysis Software Version 9.4.