A Prospective, Randomized, Multicenter Trial Comparing The

Endocrine Journal 2015, 62 (12), 1049-1057

Original A prospective, randomized, multicenter trial comparing the efficacy and safety of the concurrent use of long-acting insulin with mitiglinide or voglibose in patients with type 2 diabetes

Jang-Won Son1), In-Kyu Lee2), Jeong-taek Woo3), Sei Hyun Baik4), Hak Chul Jang5), Kwan Woo Lee6), Bong Soo Cha7), Yeon-Ah Sung8), Tae Sun Park9), Soon-Jib Yoo1)** and Kun-Ho Yoon10)*

1) Department of Internal Medicine, College of Medicine, Bucheon St. Mary’s Hospital, The Catholic University of Korea, Bucheon, Korea 2) Department of Internal Medicine, Kyungpook National University School of Medicine, Daegu, Korea 3) Department of Endocrinology and Metabolism, Research Institute of Endocrinology, School of Medicine, Kyung Hee University, Seoul, Korea 4) Department of Internal Medicine, Korea University College of Medicine, Seoul, Korea 5) Department of Internal Medicine, Seoul National University Bundang Hospital, Seongnam, Korea 6) Department of Endocrinology and Metabolism, Ajou University School of Medicine, Suwon, Korea 7) Division of Endocrinology and Metabolism, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Korea 8) Department of Internal Medicine, Ewha Womans University School of Medicine, Seoul, Korea 9) Department of Internal Medicine, Chonbuk National University Hospital, Jeonju, Korea 10)Department of Internal Medicine, College of Medicine, Seoul St. Mary’s Hospital, The Catholic University of Korea, Seoul, Korea

Abstract. This trial was conducted to compare the efficacy and safety of combination therapy with basal insulin glargine plus mitiglinide to that of basal insulin glargine plus voglibosein patients with type 2 diabetes. This was a 20-week, randomized, multicenter non-inferiority trial. Patients with HbA1c levels over 7.0% were randomly assigned to receive either mitiglinide (10 mg tid) or voglibose (0.2 mg tid) concurrent with insulin glargine for 16 weeks after a 4-week of basal insulin glargine monotherapy. The intention-to-treat population included 156 patients; 79 were placed in the mitiglinide group, and 77 were placed in the voglibose group. At 20 weeks, there was no significant difference between the mitiglinide group and the voglibose group in terms of the mean HbA1c level or the mean decrease of the HbAlc level from baseline (−0.9% [−7.5 mmol/mol] and −0.7%, [−5.3 mmol/mol] respectively). The mean fasting plasma glucose level and data of self-monitoring blood glucosewere significantly decreased from baseline to week 20 in both groups, but there was no significant difference between the two groups. The changes in the basal insulin requirements of each group were not significant. The prevalence of adverse events and the risk of hypoglycemia were similar for both groups. Combination therapy with mitiglinide plus basal insulin glargine was non-inferior to voglibose plus basal insulin glargine in terms of the effect on overall glycemic control.

Key words: Diabetes mellitus, Type 2, Mitiglinide, Voglibose

CAREFUL glycemic control is critical for protection betes [1, 2]. The goal of therapy is to normalize glu- against the disease-related complications of type 2 dia- cose levels by reducing the levels of the components

Submitted Jun. 19, 2015; Accepted Aug. 26, 2015 as EJ15-0325 of the glucose triad, which include glycated hemoglo- Released online in J-STAGE as advance publication Sep. 25, 2015 bin (HbA1c), fasting plasma glucose (FPG) and post- *Correspondence to: Kun-Ho Yoon, M.D., Ph.D., Division of prandial glucose (PPG). Increasing evidences support Endocrinology and Metabolism, Department of Internal Medicine, the importance of PPG control, in addition to FPG and College of Medicine, Seoul St. Mary’s Hospital, The Catholic University of Korea, 222 Banpo-daero, Seocho-gu, Seoul 137-040, HbAlc, for the management of type 2 diabetes [3, 4]. Korea. E-mail: [email protected] PPG was a predominant contributor to excess hyper- **Co-correspondence: Soon-Jib Yoo, M.D., Ph.D., Division of glycemia, and was recognized as an important target Endocrinology and Metabolism, Department of Internal Medicine, for treatment, especially in Asian populations with type Bucheon St. Mary’s Hospital , The Catholic University of Korea, 327 Sosa-ro, Wonmi-gu, Bucheon-si, Gyeonggi-do 420-717, Korea. 2 diabetes [5]. Many lines of evidence also indicate E-mail: [email protected] that PPG and glucose variability are major risk factors ©The Japan Endocrine Society 1050 Son et al. for cardiovascular disease and that lowered PPG lev- cal trial has compared the effect of mitiglinide with the els are associated with a reduced risk of atherosclerosis alpha-glucosidase inhibitor voglibose in combination and cardiovascular events [6-8]. Thus, many clinicians with basal insulin for the treatment of type 2 diabetes. consider several therapeutic strategies for reducing The present study was designed to compare the safety PPG excursions. and efficacy of mitiglinide versus voglibose in combi- Data from several studies suggest that initiation of nation therapy with once-daily insulin glargine for the basal insulin therapy earlier than recommended by the improvement of overall glycemic control. current standard of care would be a simple and effective method of improving FPG [9, 10]. However, for many Methods patients with type 2 diabetes, this does not improve the PPG level. Although a bolus of insulin administered at Subjects each meal is the best technique for PPG control, it is dif- Subjects with type 2 diabetes were enrolled if they ficult to maintain patient compliance with this protocol. were between 30 and 70 years old, had an HbA1c level Therefore, the addition of oral hypoglycemia agents above 7.0% (even after the administration of at least (OHAs) to basal insulin therapy as an effort to lower 2 oral antihyperglycemic drugs for 6 months prior to PPG levels is preferred over basal-bolus insulin therapy screening or after the administration of insulin glargine [11, 12]. Several classes of OHAs may be used for PPG monotherapy for at least 3 months prior to screening) control. Glinides enhance early-phase insulin release, and a body mass index (BMI) between 21 and 40 kg/ and their rapid onset and short duration of action make m2. We excluded subjects who were using types of insu- them effective at reducing PPG excursions [13]. Alpha- lin other than insulin glargine, who had an FPG over 15 glucosidase inhibitors reduce PPG by inhibiting car- mmol/L, who had C-peptide levels below 0.3 nmol/L or bohydrate digestion in the small intestine and delay- who had a history of gastrointestinal resection. In addi- ing its absorption [13]. Considering that the loss of tion, subjects with severe hepatic dysfunction, decom- first-phase insulin secretion and carbohydrate absorp- pensated cirrhosis, aspartate transaminase (AST) or tion play important roles in determining the extent of alanine transaminase (ALT) levels 2.5 times the upper PPG fluctuations, the combined use of either glinides or normal limit, unstable angina or episodes of acute myo- alpha-glucosidase inhibitors with basal insulin therapy cardial infarction within 3 months, renal failure, uncon- may provide additional PPG control, but neither reg- trolled hypertension with diastolic pressure above 110 imen is currently included in the consensus treatment mmHg (even after treatment), life-threatening diseases algorithms. Various studies have demonstrated the effi- such as cancer, severe infections, a history of drug aller- cacy of glinides or alpha-glucosidase inhibitors in com- gies, those who required the administration of oral or bination with basal insulin for the glycemic control of intravenous corticosteroids and those who were preg- patients with type 2 diabetes [14-16]. nant or lactating were not eligible for the study. This However, few studies have compared the effect of trial complied with the ethical principles defined by the each combination therapy on the glycemic control of Declaration of Helsinki and the Korean Good Clinical patients with type 2 diabetes. In one recent crossover Practice (KGCP) guidelines. The protocol was approved study, nateglinide and acarbose were reported to be by the local institutional review boards, and all subjects equally effective at controlling the mean glucose level provided written and informed consent before the initia- when combined with insulin glargine therapy, but their tion of any trial-related activities. This study is regis- effects on HbA1c were inconclusive due to the short tered ClinicalTrials.gov, number NCT00663884. duration of the study [17]. Given the different mechanisms of action between glinides and alpha-glucosidase Study design and methods inhibitors, the specific response to both regimen remains This was a multi-center, active drug-controlled, ran- to be determined. Recently, mitiglinide, a novel class of domized, open labeled, two-parallel-group-compari- glinide, was shown to be safe and effective both as a son, 20 weeks study. The clinical trial was carried out monotherapy and in combination with insulin glargine at 9 clinical trial centers from February 2008 to June [18, 19]. In two previous studies, mitiglinide and nat- 2009. Patients were divided into groups and received eglinide had similar effects on the glycemic control of either insulin glargine plus 10 mg of mitiglinide three patients with type 2 diabetes [20, 21]. To date, no clini- times daily or insulin glargine plus 0.2 mg of voglib- Mitiglinide or voglibose with glargine 1051 ose three times daily. The subjects who had agreed to control (0 - 18 points), fair control (19 - 31 points) or participate in the trial were given insulin glargine as a poor control (32 points or more). monotherapy for 4 weeks after a run-in period and were The investigators measured vital signs and body then randomly selected to receive either mitiglinide or weight and assessed the occurrence of adverse events at voglibose concurrent with the insulin glargine for 16 each visit. Routine complete blood counts, blood chem- weeks. If the subjects’ insulin glargine compliance istries and urinalysis tests were also carried out to moni- was less than 75% at week 4, they were dropped from tor drug safety. Safety was evaluated for each of the the trial. Randomization was conducted with the block patients who made at least one visit after beginning the stratified randomization method. The trial subjects study medication. The primary endpoint evaluated was visited the clinic at 4 weeks, 12 weeks and 20 weeks the change in the HbA1c level, as measured before and after registration and received insulin glargine for basal at the end of the study. The secondary endpoints evalu- insulin management throughout the entire course of ated were the mean changes in the FPG and SMBG val- the trial. The initial insulin glargine dose was 10 units ues, the mean change in the insulin dose, the proportion before bed and was injected subcutaneouslyonce daily of subjects who attained a HbA1c level less than 7.0% at the scheduled time. Adjustments to the insulin dose and the mean change in the plasma CRP level. during the trial were made according to the mean FPG value, which the patient monitored for consecutive 3 Statistical analysis days in the morning. When the mean morning FPG The minimum clinically relevant treatment differ- level exceeded 7.8mmol/L, the dose was increased by ence in HbA1c level between the groups was assumed 4 units; when the mean value was between 6.1 and 7.8 to be 0.3%, and the standard deviation was estimated mmol/L, the dose was increased by 2 units; when the at 0.7. For a study with 80% power and a 5% type mean value was between 4.4 to 6.1 mmol/L, the dose I error rate, a sample size of 68 per treatment group was maintained and when the mean value was below was required to detect the non-inferiority of mitiglinide 4.4 mmol/L, the dose was decreased by 2 units. The compared to voglibose. Assuming an overall dropout patients received education about diet/exercise therapy rate of 10% and a 1:1 randomization ratio, we enrolled and about how to manage hypoglycemia, if necessary. 80 subjects in each treatment group. The cases were After an overnight fast, the levels of FPG, HbA1c, divided into two groups for an efficacy analysis: an fasting C-peptide and high sensitivity C-reactive pro- intention-to-treat (ITT) analysis group and a per-protein (hsCRP) were measured. The serum samples col- tocol (PP) analysis group. The majority of cases for lected from the trial centers were analyzed at the central efficacy analysis were in the ITT analysis group. The laboratory. The subjects were required to measure and population that was used for the efficacy analysis (ITT record the self-measurement of blood glucose (SMBG) population) consisted of patients with at least one mea- data 3 days before the hospital visit until the day prior sured efficacy parameter among the randomized sub- to the visit. Patients measured SMBG 7 times, which jects. For the subjects who dropped out of the trial, the included the measurements before and after each meal final measurements were substituted with those from and before going to bed, and recorded their daily insu- week 16. For the subjects who did not drop out but lin doses on the log sheet. The M-value, i.e., the gly- for whom no values had been measured for the perti- cemic variability index based on the SMBG data, was nent measurement period, the values were processed calculated according to the following formula [22]: without the missing value. We tested the differences between the mitiglinide group and the voglibose group using a two-sample t-test for each continuous variable, and we tested the changes observed before and after the where W is the difference between the maximum and administration of mitiglinide or voglibose within the minimum blood glucose values (mg/dL) during the groups using paired t-tests. The between-group differ- period and n is the number of blood glucose measure- ences in achieving the HbA1c target level were tested ments taken. using the Chi-squared or Fisher’s exact tests. The The M-value was used to evaluate the adequacy of statistical analysis was carried out using SAS 9.1 for blood glucose control via the monitoring of plasma glu- Windows (SAS Institute Inc., Cary, NC). P values less cose levels. Patients were categorized as having good than 0.05 were considered to be statistically significant. 1052 Son et al.

Fig. 1 Disposition of subjects in the trial

Results Fig. 2. Both groups showed significant decreases in FPG levels during the study period. The FPG levels We screened 167 subjects and included 156 sub- decreased by 1.25 ± 3.55 mmol/L in the mitiglinide jects in the ITT population, 79 in the mitiglinide group group and by 0.94 ± 3.05 mmol/L in the voglibose and 77 in the voglibose group. Within the ITT analy- group. The SMBG measurements of all the other time sis group, 105 patients (mitiglinide group: 59; voglib- points at week 20 also significantly decreased com- ose group: 46) were selected for the PP-analysis group pared with baseline within each mitiglinide (Fig. 2B) (Fig. 1). There were no statistically significant differ- and voglibose groups (Fig. 2C). The mitiglinide group ences in the clinical characteristics of the subjects in exhibited a greater decrease, but the difference between each group besides age (p = 0.0011) and a history of groups was not significant. smoking (p = 0.0430) in Table 1. As shown in Table 2, the M-values that were calculated from the SMBG measurements indicated that Efficacy the mitiglinide group scored 41.4 ± 29.4 points prior At the conclusion of the study, the HbA1c levels had to combination therapy and 23.8 ± 19.0 points after 20 decreased by 0.9 ± 1.0% (7.5 ± 8.6 mmol/mol) in the weeks, while the voglibose group scored 49.6 ± 34.4 mitiglinide group and by 0.7 ± 1.2% (5.3±10.8mmol/ points at the beginning of the study and 28.7 ± 25.8 mol) in the voglibose group (Fig. 2A). The decreases points after 20 weeks. These decreases were signifi- were significant in both groups, and the 95% one-sided cant in both groups, but the difference in the decreased confidence interval for the mean change between the M-values between the groups was not significant. After two groups was -∞ to 0.09%. This value did not exceed the completion of the study protocol, there were 2 the 0.3% non-inferiority margin and hence demon- (2.5%) subjects in the mitiglinide group and 1 (1.3%) strated the non-inferiority of mitiglinide compared to subject in the voglibose group with HbA1c levels below voglibose. The change in the daily profiles of SMBG 6.5% (48 mmol/mol), whereas there were 13 (16.5%) levels from baseline to the end of study are shown in subjects in the mitiglinide group and 5 (6.5%) subjects Mitiglinide or voglibose with glargine 1053

Table 1 Baseline characteristics of the subjects in the intention- to-treat population Mitiglinide Voglibose Variables group group (n = 79) (n = 77) Age (yrs) 51.9 ± 8.3 56.1 ± 7.6 Men, n (%) 45 (57.0) 47 (61.0) Duration of diabetes (yrs) 10.6 ± 5.5 12.2 ± 6.1 Smokers, n (%) 12 (15.2) 22 (28.6) Bodyweight (kg) 67.5 ± 10.2 67.0 ± 10.1 BMI (kg/m2) 25.1 ± 2.5 25.0 ± 2.8 HbA1c NGSP (%) 8.9 ±1.0 9.1 ± 1.0 IFCC (mmol/mol) 75 ± 9.5 77 ± 9.6 Fasting plasma glucose (mmol/L) 7.2 ± 3.2 7.2 ± 2.6 Fasting C-peptide (nmol/L) 0.85 ± 0.34 0.87 ± 0.53 Obese Obese n 34 28 (BMI>25) C-peptide level 0.98 ± 0.41 1.14 ± 0.77 Non-obese n 45 49 (BMI≤25) C-peptide level 0.76 ± 0.25 0.72 ± 0.21 Duration of DM n 69 70 ≥ 5yrs C-peptide level 0.86 ± 0.35 0.86 ± 0.55 n 10 7 < 5yrs C-peptide level 0.82 ± 0.24 1.03 ± 0.30 Total cholesterol (mmol/L) 4.5 ± 1.0 4.2 ± 0.9 Triglycerides (mmol/L) 1.7 ± 1.2 1.6 ± 1.0 HDL-cholesterol (mmol/L) 1.2 ± 0.2 1.2 ± 0.2 LDL-cholesterol (mmol/L) 2.4 ± 0.8 2.2 ± 0.7 Fig. 2 Changes in plasma glycated hemoglobin (HbA1c) level C-reactive protein (mg/dL) 0.19 ± 0.32 0.17 ± 0.55 after combination therapy. (A) Changes in daily profiles of glucose levels at 7-points from baseline to week 20 History of sulfonylurea 78 (98.7) 74 (96.1) in patients treated with mitiglinide (B) or voglibose (C). medication, n (%) BB, before breakfast; AB, 2 hours after breakfast; BL, History of insulin agent 16 (20.3) 14 (18.2) before lunch; AL, 2 hours after lunch; BD, before dinner; administration, n (%) AD, 2 hours after dinner; BS, before sleep. * p < 0.05; Data are expressed as the means ± SD or percentage. Change from baseline in the voglibose group with HbA1c levels below 7% (53 group demonstrated a smaller increase in insulin dose, mmol/mol). Although the rate of individuals achieving the difference between the two groups was not signifi- the target HbA1c level was higher for the mitiglinide cant. The hsCRP level is commonly used as an inflam- group, the between-group difference in these rates was matory response index, and some reports indicate that it not significant (p = 0.0515). Both groups received can also be related to insulin resistance. Therefore, the significantly increased insulin doses during the study changes in hsCRP level before and after the administra- period, as the mitiglinide group reported a 3.3 ± 10.3 tion of the investigated drugs was evaluated; the miti- unit increase and the voglibose group reported a 6.0 ± glinide group experienced a 0.06 mg/dL decrease, and 10.4 unit increase. Once again, although the mitiglinide the voglibose group was found to have a 0.04 mg/dL 1054 Son et al.

Table 2 Changes in glycemic control over the 16 weeks of treatment Table 3 Adverse events by mitiglinide and voglibose Mitiglinide Voglibose Mitiglinide Voglibose group group p value group group p value (n = 79) (n = 77) (n = 79) (n = 81) HbA1c (%) Serious adverse events (SAE) 1 (1.3) 5 (6.2) 0.2101* Baseline 9.0 ± 1.1 9.2 ± 1.1 Liver and biliary disorders End of study 8.0 ± 1.1 8.4 ± 1.3 Cholecystitis 0 (0.00) 1 (0.01) Change from baseline -0.9 ± 1.0* -0.7 ± 1.2*,† 0.2445 Secondary terms Fasting plasma glucose (mmol/L) Intervertebral disc protrusion 1 (0.01) 0 (0.00) Baseline 7.2 ± 3.2 7.2 ± 2.6 Burn 0 (0.00) 1 (0.01) End of study 5.9 ± 1.7 6.2 ± 2.3 Respiratory system disorders * * Change from baseline -1.2 ± 3.5 -0.9 ± 3.0 0.5559 Bronchopneumonia 0 (0.00) 1 (0.01) M-value Throat infection 0 (0.00) 1 (0.01) Baseline 41.4 ± 29.4 49.6 ± 34.4 Neoplasm End of study 23.8 ± 19.0 28.7 ± 25.8 Gastric cancer, stage 0 0 (0.00) 1 (0.01) * * Change from baseline -18.5 ± 28.1 -20.8 ± 33.4 0.6514 Gastrointestinal adverse events 4 (5.1) 8 (9.9) 0.3693* Daily insulin dose (U) Any hypoglycemia 1 (1.3) 4 (4.9) 0.3673* Baseline 27.0 ± 12.4 26.7 ± 9.7 Adverse drug reaction 11 (13.9) 10 (12.3) 0.7675† End of study 30.4 ± 13.8 32.7 ± 14.4 Liver and biliary disorders 0 (0.0) 1 (1.2) Change from baseline 3.3 ± 10.3 6.0 ± 10.4 0.1069 Increased AST 0 (0.0) 1 (1.2) C-reactive protein (mg/dL) Increased ALT 0 (0.0) 1 (1.2) Baseline 0.2 ± 0.3 0.2 ± 0.6 Metabolic and nutritional disorders 1 (1.3) 2 (2.5) End of study 0.1 ± 0.1 0.1 ± 0.2 Hyperlipidemia 1 (1.3) 0 (0.0) Change from baseline -0.06 ± 0.3 -0.04 ± 0.5 0.6898 Hypoglycemia 0 (0.0) 1 (1.2) Body Weight Hypoglycemia, unperceived 0 (0.0) 2 (2.5) Baseline 67.6 ± 10.3 67.1 ± 10.2 Visual disorders 1 (1.3) 0 (0.0) End of study 68.5 ± 10.1 67.1 ± 10.4 Retinopathy, diabetes 1 (1.3) 0 (0.0) Change from baseline 0.93 ± 2.6* 0.01 ± 2.0 0.0137 Gastrointestinal disorders 1 (1.3) 0 (0.0) Achievement of treatment goal Retching 1 (1.3) 0 (0.0) HbA1c < 7.0%, n (%) 13 (16.5) 5 (6.5) 0.0515 Systemic-general 4 (5.1) 1 (1.2) Data are expressed as the means ± SD. *p < 0.05 was calculated Tremor 0 (0.0) 1 (1.2) by using paired t-test. †The 95% one-sided confidence interval of the mean-difference of the two groups was (-∞ to 0.09%), which Inertia 2 (2.5) 0 (0.0) did not exceed the 0.3% non-inferiority margin. Weakness, systemic 2 (2.5) 0 (0.0) Powerlessness 1 (1.3) 0 (0.0) Neuropsychologic disorders 1 (1.3) 2 (2.5) decrease. Neither the difference between the pre- and Abnormal hunger 1 (1.3) 2 (2.5) post-treatment levels within each group nor the differ- Central and peripheral nerve 5 (6.3) 4 (4.9) ence between the groups was significant. Myokymia 1 (1.3) 0 (0.0) The drug efficacy was also evaluated in the Headache 0 (0.0) 1 (1.2) PP-analysis group. Most of the variables evaluated were found to be similar to those of the ITT-analysis Tremor 1 (1.3) 0 (0.0) group, and no data for any of the categories were signif- Tremor, minor 1 (1.3) 0 (0.0) icantly different between the two groups. Tremor, limbs 1 (1.3) 0 (0.0) Vertigo 2 (2.5) 3 (3.7) Safety Skin disorder 2 (2.5) 1 (1.2) A safety evaluation was carried out for 160 patients Perspiration 2 (2.5) 0 (0.0) (mitiglinide group: 79; voglibose group: 81) in Table Rosacea 0 (0.0) 1 (1.2) 3. There was at least one adverse event experienced *Fisher’s exact test. †Pearson’s chi-square test. Mitiglinide or voglibose with glargine 1055 by 35 patients (44.30%) in the mitiglinide group and insulin glargine for glycemic control. 41 patients (50.62%) in the voglibose group, but the The primary endpoint of this study was the reduc- between-group difference was not significant (p = tion of the HbA1c level after 16 weeks of treatment. 0.4240). There were 5 subjects (3.13%) who experi- For this measure, the administration of mitiglinide with enced hypoglycemic adverse events; 1 (1.26%) of these insulin glargine was non-inferior to combination ther- were in the mitiglinide group and 4 (4.94%) were in apy with voglibose. At the end of the study, there was the voglibose group, and there was no significant differ- no significant between-group difference in the achieve- ence between groups (p = 0.3673). Adverse drug reac- ment rate of the HbA1c goal. However, more subjects tions were showed in 21 subjects; 11 (13.9%) of these in the mitiglinide group exhibited HbA1c decreases were in the mitiglinide group and 10 (12.3%) of these greater than 0.5% [mitiglinide group: 55 (69.9%); were in the voglibose group, and there was not showing voglibose group: 44 (57.1%)]. The FPG and SMBG any difference of treatment groups (p = 0.7675). Six secondary endpoint values were also similar between serious adverse events were reported in 6 subjects of the mitiglinide group and the voglibose group, although which one subject was in the mitiglinide group and 5 the mitiglinide group required a smaller increase in the subjects were in the voglibose group, but the events did insulin dose during basal insulin therapy. Both reg- not exhibit a causal relationship with the investigated imens were associated with similar frequencies of drug treatments (p = 0.2101). There was no statistically adverse events, including hypoglycemic events. These significant between-group difference in the assignable findings correspond well to those of similar compara- variations for any laboratory test. For the differences tive studies that have reported the efficacy of glinide between the groups in terms of vital sign measurement, and acarbose combined with insulin glargine therapy body weight and BMI, the mitiglinide group showed a for glycemic control [17, 25]. significant increase in body weight (gain of 0.9kg). A recent study, which evaluated the use of insulin glargine in combination with either repaglinide or acar- Discussion bose, found that both combination therapies decreased HbA1c levels by approximately 3.0% (9 mmol/mol) The objective of this trial was to comparatively eval- and were equally effective at reducing FPG and PPG uate the efficacy and safety of adding mitiglinide or excursions [25]. However, the HbA1c, FPG and PPG voglibose to the basal insulin glargine treatment reg- levels remained above the recommended level at the imen of patients with type 2 diabetes. The primary end of the study period. These results were likely a and secondary efficacy endpoints assessed were the result of the small number of study participants (each changes in the HbA1c, FPG, and SMBG values, the group consisted of 20 patients), the high HbA1c lev- change in the insulin dose throughout treatment and the els (11%) at the start of the trial and the absence of an proportion of subjects who attained an HbA1c level < insulin pre-treatment period. In addition, these stud- 7.0% (53 mmol/mol). Mitiglinide stimulates insulin ies reported that weight gain and severe hypoglycemic secretion by blocking the KATP channel in pancreatic events were more prevalent in the repaglinide group. beta cells, which thereby induces an influx of excess In our study, the mitiglinide group also experienced calcium. This drug exhibits an extremely rapid onset weight gain of 0.9kg, but there was no between-group and short duration for lowering the plasma glucose difference in the number of hypoglycemic events. level, and it is therefore effective at reducing postpran- Another study compared the efficacy and safety of nat- dial hyperglycemia and carries a lower risk of hypo- eglinide or acarbose when combined with basal insu- glycemia than sulfonylurea [23]. Mitiglinide has been lin for the control of PPG after the FPG had been opti- widely used alone or in combination with other OHA mized with insulin glargine prior to randomization and basal insulin therapies for the glycemic control of [18]. This study also reported no significant differ- type 2 diabetes. Voglibose is a well-established sec- ences in the mean glucose level, PPG excursions or the ond-generation alpha-glucosidase inhibitor. In Asia, mean average glucose excursions between the two reg- because of predominant high PPG, use of glinides and imens, and these results were similar to those of our alpha-glucosidase is quite popular than western coun- study. These authors did suggest that beta cell function tries [24]. However, no studies have compared the may be an important factor for predicting the response effects of mitiglinide or voglibose in combination with to treatment. Similarly, in our study, the concurrent use 1056 Son et al. of mitiglinide with insulin glargine was shown to be reported that patients who responded to the combina- more effective for subjects with C-peptide levels above tion of mitiglinide with insulin glargine were younger, 1.0 nmol/L. For subjects with C-peptide levels above more obese and had higher endogenous insulin secre- 1.0 nmol/L, the HbA1c levels were decreased by 1.2% tion capacity than non-responders [27]. (10.8 mmol/mol) in the mitiglinide group. In the vogli- We found no significant differences between the two bose group, there was no significant difference in the groups in terms of the frequency of adverse events, HbA1c level based on C-peptide level. the frequency of hypoglycemic events, or abnormal Moreover, considering the different mechanisms of laboratory test and physical examination results. The action of mitiglinide and voglibose, we carried out sub- majority of the events observed were of mild to moder- group analyses to verify the effectiveness of each drug. ate intensity. The limitations of this study included the By examining changes in HbA1c levels in regards to use of an open-label design and the fact that the anal- body weight before and after drug treatment, we found ysis was restricted to Asian patients. Patient aware- that patients with body weights greater than 80 kg in ness as to thetreatment received may have influenced the mitiglinide group had a 1.2% (10.8 mmol/mol) glycemic control-related lifestyle modifications and decrease in their HbA1c levels and those patients in the the reporting of adverse events during the study period. voglibose group had a 0.5% (3.1 mmol/mol) decrease. In addition, number of subjects who dropped out the The decrease seen for the mitiglinide group was signif- study and excluded from the PP analysis in voglibose icant, but the decrease of the voglibose group was not. (n=35/81) is around 1.7 times more than that in miti- A similar tendency was found when the results ana- glinide (n=20/79). It is mainly due to the subject devi- lyzed by BMI were examined. In the BMI subgroups ation rather than investigator selection bias. However, of over 29, the mitiglinide group showed a statistically this protocol deviation does not affect the overall results significant HbA1c decrease (−1.5% [−14 mmol/mol],p in a PP analysis compared to the analysis of changes = 0.0196) unlike the voglibose group (−0.0% [0 mmol/ within andbetween the treatment groups for efficacy mol], p = 0.9063). Recently, one study showed that a variables using an ITT analysis. In conclusion, this greater effect of nateglinide was observed in patients study confirmed the non-inferiority of the concomitant with a more severe insulin resistance [26]. Obese type use of insulin glargine with mitiglinide compared to 2 diabetes patients were characterized by an increased voglibose. Various evaluations verified that both regi- insulin resistance rather than the loss of early phase of mens helped to control and improve glycemic control insulin secretion. In our study, we assumed that miti- in patients with type 2 diabetes. Furthermore, the use glinide improve PPG level via amelioration of insu- of mitiglinide with insulin glargine is suggested as a lin resistance as well as augmentation of early insu- more effective therapy for overall glycemic control. lin response in obese type 2 diabetes patients. When changes in HbA1c levels were analyzed according to Conflict of Interest the duration of diabetes, only patients with a duration of less than five years in the mitiglinide group exhibited The authors have no conflicts of interest to declare. significant decreases in their HbA1c levels upon drug therapy. It was anticipated that the concurrent use of Acknowledgement mitiglinide with insulin glargine would be more effective for subjects with greater body weight, a shorter This work was supported by JW Pharmaceutical diabetic duration, relatively normal pancreatic beta Corporation, Korea. This study is registered cell function and high C-peptide levels. These find- ClinicalTrials.gov, number NCT00663884. ings were consistent with an earlier study, which had

References 1. UK Prospective Diabetes Study (UKPDS) Group tes (UKPDS 33). Lancet 352: 837-853. (1998) Intensive blood-glucose control with sulphonyl- 2. Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, ureas or insulin compared with conventional treatment et al. (1995) Intensive insulin therapy prevents the pro- and risk of complications in patients with type 2 diabe- gression of diabetic microvascular complications in Mitiglinide or voglibose with glargine 1057

Japanese patients with non-insulin-dependent diabe- Monnier L, et al. (2005) Postprandial glucose regula- tes mellitus: a randomized prospective 6-year study. tion: new data and new implications. Clin Ther 27: S42- Diabetes Res Clin Pract 28: 103-117. 56. 3. Monnier L, Lapinski H, Colette C (2003) Contributions 17. Kim MK, Suk JH, Kwon MJ, Chung HS, Yoon CS, et al. of fasting and postprandial plasma glucose increments (2011) Nateglinide and acarbose for postprandial glu- to the overall diurnal hyperglycemia of type 2 diabetic cose control after optimizing fasting glucose with insu- patients: variations with increasing levels of HbA(1c). lin glargine in patients with type 2 diabetes. Diabetes Diabetes Care 26: 881-885. Res Clin Pract 92: 322-328. 4. El-Kebbi IM, Ziemer DC, Cook CB, Gallina DL, 18. Malaisse WJ (2008) Mitiglinide: a rapid- and short- Barnes CS, et al. (2004) Utility of casual postprandial acting non-sulfonylurea insulinotropic agent for the glucose levels in type 2 diabetes management. Diabetes treatment of type 2 diabetic patients. Expert Opin Care 27: 335-339. Pharmacother 9: 2691-2698. 5. Wang JS, Tu ST, Lee IT, Lin SD, Lin SY, et al. (2011) 19. Kumashiro N, Yoshihara T, Kanazawa Y, Shimizu T, Contribution of postprandial glucose to excess hyper- Watada H, et al. (2007) Long-term effect of combination glycaemia in Asian type 2 diabetic patients using con- therapy with mitiglinide and once daily insulin glargine tinuous glucose monitoring. Diabetes Metab Res Rev in patients who were successfully switched from inten- 27: 79-84. sive insulin therapy in short-term study. Endocr J 54: 6. Raz I, Wilson PW, Strojek K, Kowalska I, Bozikov V, 163-166. et al. (2009) Effects of prandial versus fasting glyce- 20. Gao X (2009) Multicentre, double-blind, randomized mia on cardiovascular outcomes in type 2 diabetes: the study of mitiglinide compared with nateglinide in type HEART2D trial. Diabetes Care 32: 381-386. 2 diabetes mellitus patients in China. J Int Med Res 37: 7. Ceriello A, Cavarape A, Martinelli L, Da Ros R, Marra 812-821. G, et al. (2004) The post-prandial state in Type 2 dia- 21. Li L, Yang M, Li Z, Yan X, Guo H, et al. (2012) Efficacy betes and endothelial dysfunction: effects of insulin and safety of mitiglinide versus nateglinide in newly aspart. Diabet Med 21: 171-175. diagnose patients with type 2 diabetes mellitus: a ran- 8. Esposito K, Giugliano D, Nappo F, Marfella R (2004) domized double blind trial. Diabetes Obes Metab 14: Regression of carotid atherosclerosis by control of post- 187-189. prandial hyperglycemia in type 2 diabetes mellitus. 22. Schlichtkrull J, Munck O, Jersild M (1965) The m-value, Circulation 110: 214-219. an index of blood-sugar control in diabetics. Acta Med 9. Li Y, Xu W, Liao Z, Yao B, Chen X, et al. (2004) Scand 177: 95-102. Induction of long-term glycemic control in newly 23. Sunaga Y, Gonoi T, Shibasaki T, Ichikawa K, Kusama diagnosed type 2 diabetic patients is associated with H, et al. (2001) The effects of mitiglinide (KAD-1229), improvement of beta-cell function. Diabetes Care 27: a new anti-diabetic drug, on ATP-sensitive K+ channels 2597-2602. and insulin secretion: comparison with the sulfonyl- 10. Ryan EA, Imes S, Wallace C (2004) Short-term inten- ureas and nateglinide. Eur J Pharmacol 431: 119-125. sive insulin therapy in newly diagnosed type 2 diabetes. 24. International Diabetes Federation (2012) Global Diabetes Care 27: 1028-1032. Guideline for Type 2 Diabetes [accessed Jan 2014] 11. Yki-Jarvinen H (2001) Combination therapies with Available from: http://www.idf.org/sites/default/files/ insulin in type 2 diabetes. Diabetes Care 24: 758-767. IDF%20T2DM%20Guideline.pdf 12. Riddle MC (2008) Combined therapy with insulin plus 25. Duran C, Tuncel E, Ersoy C, Ercan I, Selimoglu H, et oral agents: is there any advantage? An argument in al. (2009) The investigation of the efficacy of insulin favor. Diabetes Care 31 Suppl 2: S125-130. glargine onglycemic control when combined with either 13. Inzucchi SE (2002) Oral antihyperglycemic therapy for repaglinide oracarbose in obese type 2 diabetic patients. type 2 diabetes: scientific review. JAMA 287: 360-372. J Endocrinol Invest 32:69–73. 14. Dashora UK, Sibal L, Ashwell SG, Home PD (2007) 26. Shiba T (2003) Improvement of insulin resistance by a Insulin glargine in combination with nateglinide in peo- new insulin secretagogue, nateglinide-analysis based on ple with Type 2 diabetes: a randomized placebo-con- the homeostasis model. Diabetes Res Clin Pract 62:87- trolled trial. Diabet Med 24: 344-349. 94. 15. Hwu CM, Ho LT, Fuh MM, Siu SC, Sutanegara D, et 27. Yoshihara T, Kumashiro N, Kanazawa Y, Mita T, al. (2003) Acarbose improves glycemic control in insu- Sakurai Y, et al. (2006) Therapeutic efficacy of miti- lin-treated Asian type 2 diabetic patients: results from glinide combined with once daily insulin glargine after a multinational, placebo-controlled study. Diabetes Res switching from multiple daily insulin regimen of aspart Clin Pract 60: 111-118. insulin and glargine in patients with type 2 diabetes 16. Leiter LA, Ceriello A, Davidson JA, Hanefeld M, mellitus. Endocr J 53: 67-72.