Study design and setting
The study was a 52-week, multicentered, randomized and positive-controlled clinical trial which included a 4-week run-in period. Between Nov 25, 2011, and Jun 27, 2013, a total of 450 subjects were enrolled from 9 clinical research sites after run-in period (Fig. 1A). Informed consent was obtained from every study subject. The research protocol was approved by the Guang’anmen Hospital Medical Ethics Commission in Beijing, China. The clinical trial registration number of this study is NCT01471275 (www.ClinicalTrials.gov). The study was conducted in accordance with the Declaration of Helsinki guidelines on good clinical practices [22].
Patients
Inclusion criteria. The early type 2 diabetic status of these patients was confirmed according to the diagnosis standard [23]. Subjects, 18-70 years old, had a waistline male ≥ 90cm, female ≥ 80cm. After the initial screening, subjects entered a 4-week run-in period with diet control and programmed daily exercise. Then, those subjects with sustained HbA1c≥7.0% and FPG level between 7.0 and 13.9 mmol/L, or 2-h postprandial plasma glucose (2-h PG) ≥11.1 mmol/L, triglyceride (TG) between 1.7 and 5.65 mmol/L were enrolled. A consent form was signed by all subjects prior to enrollment.
Exclusion criteria. Any patient with the following conditions were excluded from participation of the trial: those who had used insulin therapy; had been treated for diabetes for over 3 months by conventional medications, physical therapy, psychological therapy, herbal medicines or dietary supplements; had been treated with antidiabetic and lipid-lowering drugs 1 month prior to screening; had complications of diabetes; had serious heart, lung, liver, kidney, brain related primary diseases or complications; had uncontrolled hypertension (SBP ≥ 160 mmHg and/or DBP ≥100 mmHg); had diabetic ketoacidosis or serious infections within 1 month; had mental disorders; pregnant females, or those planning to be pregnant; breast feeding; allergic to Chinese herbal medicines; participating in other clinical trials within 1 month prior to screening; alcoholism, taking antipsychotic agents or substance abuse or dependence; had factors that may affect trial execution based on investigators’ judgment, such as changeable working and living environments that may lead to withdrawal from the trial; had unstable antihypertension effects during drug administration; or taking weight-loss medicines; had hepatic and renal dysfunctions [aspartate aminotransferase (AST), alanine aminotransferase (ALT) more than two times higher than the normal upper limit and serum creatinine above normal upper limit]; insensitive to hypoglycemia.
Withdrawal criteria. Patients who met the following criteria were asked to withdraw from the trial: those who experienced serious diabetic complications (such as diabetic ketoacidosis, hyperosmolar nonketotic syndrome, lactic acidosis, hypoglycemic coma); had poor compliance, with test medication use less than 80% of or more than 120% of the prescribed dose; had violated protocol, such as taking another oral hypoglycemic agent.
Intervention
Study Medication. JTTZF (manufactured by Tianjiang Pharmaceutical, Jiangyin, China) and metformin (by the Lilinghengtai Pharmaceutical, Beijing, China; batch number: 110508) were used. The JTTZF, prepared in formula granules, consists of eight Chinese medicinal herbs, including rhizoma anemarrhenae (batch number: 1110318), monascus (batch number: 1209335), aloe (batch number: 1110316), coptis chinensis (batch number: 1110317), balsam pear (batch number: 1110322), salvia miltiorrhiza (batch number: 1110323), schisandra chinensis (batch number: 1110320) and rhizoma zingiberis (batch number: 1110319). The manufacturer was not involved in the design and analysis of this study.
The subjects were randomly allocated to receive either JTTZF or metformin for 48 consecutive weeks. For JTTZF group, each subject ingested 30 g with warm water, two times daily with meals. For metformin group, subjects ingested 0.25 g with warm water, three times daily after each meal.
Chemical Analysis of JTTZF. The formulation of JTTZF was prepared according to the 2005 Edition of Chinese Pharmacopoeia. The chemical composition of JTTZF was measured by using an ultra-performance liquid chromatography/mass (UPLC/MS) method. An example chromatographic fingerprint is shown in Figure 2. Chemical structures of eight major compounds (mangiferin, coptisine chloride, jatrorrhizine, lithospermic acid B, aloin, berberine, palmatine, lovastatin) in the finished dosage are also shown. These compounds were also used as the quality control markers in the JTTZF formulation.
Adjustments of Drugs and Dosages. The medication remained stable during the first 24-week trial period for all the subjects. If HbA1c was over 7% at 24-week, subjects should add gliclazide sustained-release tablets once per day; additionally, if their TG was over 5.65 mmol/L, add bezafibrate twice per day.
Co-administration with Other Drugs. Subjects were not allowed to take weight-loss medicines and other antidiabetic drugs throughout the study. For those subjects who had to take other drugs related to other diseases during the trial period, the dosage of these medications should be recorded in combination therapy table.
Test Medication Delivery. There were thirteen visits during the trial period in every 4weeks. During each visit, physical and biochemical tests were performed based on the trial design. Each patient was given a 4-week drug supply. Unused testing medications for the period before each visit were re-collected and counted for compliance check.
Randomization
A stratified, block randomization method was conducted by the study center. Study drugs were packed and numbered according to the random coding form and randomly allocated to each research site using concealed opaque envelopes. These envelopes and case report forms were not collected until the end of the trial. Study drugs were provided based on the assigned numbers, which were determined according to the visit sequence and study drug number sequence, and remained unchanged throughout the trial. The only basis of drug distribution was the unique drug number. Independent statisticians performed the data analysis.
2.5 Endpoints and assessment
During this 48-week period, subjects were assessed at each 4weeks. In each session, subjects were asked if there were any adverse events. All subjects received a symptom assessment, physical examination, and the compliance of the test drug administration. FPG, waist circumstance, body weight and BMI were monitored. The HbA1c, 2-h postprandial PG (2-h PG), triglycerides (TG), total cholesterol (TC), high-density lipoprotein (HDL) and low-density lipoprotein (LDL), fasting insulin were measured at week 0, 12, 24, 36 and 48.
Primary Endpoints. The primary endpoints were change in HbA1c from baseline to 48 weeks and proportion of patients achieving HbA1c<7.0, <6.5, and <6.0 at week 48. The HbA1c level was measured in a central laboratory (Guang’anmen Hospital, China Academy of Chinese Medical Sciences) using an ADAMSPMALCHA-8160 automated HbA1c analyzer (Japan).
Secondary Endpoints. The secondary endpoints were changes in the FPG, 2hPG, fasting insulin, lipids, waist circumstance and the body weight between the two groups. Fingertip blood was collected at each visit and FPG was measured using a blood glucose analyzer (ACCU-CHEK Active meter, Roche Diagnostics, Indianapolis, USA). The 2hPG level was measured at each 12 weeks (Olympus AU2700 Analyzer, Olympus Life and Material Science, Europa GmbH, Hamburg, Germany). An immunonephelometry method was used to measure lipids (COBAS Integra 400 Plus System, Roche Diagnostics Ltd. Basel, Switzerland). Immunoreactive insulin was measured by electrochemiluminescence immuno assay (Roche Diagnostics Ltd, Basel, Switzerland), and the homeostatic model assessment (HOMA) was performed to quantify insulin resistance (HOMA-IR) and β-cell function (HOMA-β). Percentages of patients achieving TG <1·7 mmol/L, TC <5·7 mmol/L, and body mass index (BMI) were also evaluated.
Safety Evaluation. Vital signs were collected, hepatic functions (ALT and AST) and renal function (serum creatinine and BUN) were measured at each 4 weeks. Routine blood tests, urine tests, stool tests (including occult blood test) and ECG were determined at week 0, 12, 24, 36 and 48. Adverse events were recorded immediately after reporting.
Statistical Analysis
Data entry was completed twice by two staff members using Epidata software. Numerical variables were summarized as means and standard deviations (SD), mean differences and proportions of change between baseline and the 48-week follow-ups. For between-group comparison, t-test or Wilcoxon rank sum test was performed. When examining the central effects and covariates, the analysis of covariance was used. χ2 test or Fisher's exact test were used to compare the incidence of adverse events between the two groups. The level of statistical significance was set at P < 0.05.