E-dmj.org

Original Article
Clinical Care/Education
Diabetes Metab J 2013;37:465-474
http://dx.doi.org/10.4093/dmj.2013.37.6.465 D I A B E T E S & M E T A B O L I S M J O U R N A L
Glycemic Effectiveness of Metformin-Based Dual-Combination Therapies with Sulphonylurea, Pioglitazone, or DPP4-Inhibitor in Drug-Naïve Korean Type 2 Diabetic Patients Young Ki Lee1,2, Sun Ok Song2,3, Kwang Joon Kim1, Yongin Cho1, Younjeong Choi1, Yujung Yun1, Byung-Wan Lee1, Eun-Seok Kang1, Bong Soo Cha1, Hyun Chul Lee11Division of Endocrinology and Metabolism, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, 2Department of Medicine, The Graduate School of Yonsei University, Seoul, 3Division of Endocrinology, Department of Internal Medicine, National Health Insurance Corporation Ilsan Hospital, Goyang, Korea Background: This study compared the glycemic effectiveness of three metformin-based dual therapies according to baseline
hemoglobin A1c (HbA1c) to evaluate the appropriateness of the guideline enforced by the National Health Insurance Corpora- tion of Korea for initial medication of type 2 diabetes (T2D).
Methods: This prospective observational study was conducted across 24 weeks for drug-naïve Korean T2D patients with HbA1c
greater than 7.5%. Subjects were first divided into three groups based on the agent combined with metformin (group 1, gliclazide- modified release or glimepiride; group 2, pioglitazone; group 3, sitagliptin). Subjects were also classified into three categories ac- cording to baseline HbA1c (category I, 7.5%≤HbA1c<9.0%; category II, 9.0%≤HbA1c<11.0%; category III, 11.0%≤HbA1c). Results: Among 116 subjects, 99 subjects completed the study, with 88 subjects maintaining the initial medication. While each
of the metformin-based dual therapies showed a significant decrease in HbA1c (group 1, 8.9% to 6.4%; group 2, 9.0% to 6.6%; group 3, 9.3% to 6.3%; P<0.001 for each), there was no significant difference in the magnitude of HbA1c change among the groups. While the three HbA1c categories showed significantly different baseline HbA1c levels (8.2% vs. 9.9% vs. 11.9%; P<0.001), endpoint HbA1c was not different (6.4% vs. 6.6% vs. 6.0%; P=0.051).
Conclusion: The three dual therapies using a combination of metformin and either sulfonylurea, pioglitazone, or sitagliptin
showed similar glycemic effectiveness among drug-naïve Korean T2D patients. In addition, these regimens were similarly effec- tive across a wide range of baseline HbA1c levels.
Keywords: Diabetes mellitus, type 2; Metformin; Pioglitazone; Sitagliptin; Sulphonylurea
INTRODUCTION
is essential y based on the results of controlled clinical trials, such as the Kumamoto study and UK Prospective Diabetes In reducing microvascular and macrovascular diabetic com- Study, which are prospective randomized studies including a plications, there has been little controversy on the need for ear- large number of Asian and Western subjects, respectively [1-3]. ly intensive glycemic control in subjects with newly detected However, optimal or recommended regimens regarding the se-type 2 diabetes (T2D) since the late 1990s [1]. This consensus lection of hypoglycemic agents to effectively and safely achieve This is an Open Access article distributed under the terms of the Creative Commons At- tribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) Department of Internal Medicine, Yonsei University College of Medicine, which permits unrestricted non-commercial use, distribution, and reproduction in any 50 Yonsei-ro, Seodaemun-gu, Seoul 120-749, Korea medium, provided the original work is properly cited.
Received: Apr. 14, 2013; Accepted: Jun. 3, 2013 Copyright 2013 Korean Diabetes Association good glycemic status have differed slightly between several decisions according to usual practice. Participants were exam-guidelines [4-6]. In 2011, the Korean Diabetes Association ined every 12 to 13 weeks for 24 weeks after the initiation of recommended Clinical Practice Guidelines for T2D in Korea metformin-based dual combination therapy with OHAs.
[5]. This guideline recommended lifestyle interventions with We included only drug-naïve T2D patients with an initial metformin as an initial treatment regimen. In addition, initial HbA1c level ≥7.5% who were first-time visitors to the Sever-treatment with a combination of oral hypoglycemic agents ance Diabetes Center. In the final analysis, subjects were ex-(OHAs) or insulin was also recommended at a hemoglobin cluded if they had a recent (≤6 months) history of major car-A1c (HbA1c) level greater than 8.0% at the time of T2D diag- diovascular event, including myocardial infarction, unstable nosis. In the same year, the National Health Insurance Corpo- angina, moderate to severe congestive heart failure, and/or ration (NHIC) established guidelines to enforce metformin- stroke. In addition, patients with a current hepatic (aspartate preferred monotherapy as a general initial treatment regimen aminotransferase [AST], and alanine aminotransferase [ALT] and metformin-based dual therapies with sulphonylurea (SU), >2.5-fold the upper normal limit), renal (serum creatinine pioglitazone, or DPP4-inhibitor as an initial regimen at an >1.5 mg/dL in men, >1.4 mg/dL in women), hematologic, or HbA1c level greater than 7.5%. Recently, Yoon et al. [7] re- gastrointestinal disease or those that had undergone systemic ported a reduction in HbA1c level after conducting a double- corticosteroid treatment in the previous 12 weeks were exclud- blind, randomized controlled study over a 48-week period on ed. Subjects recruited for the study were classified into three the efficacy of glimepiride, metformin, and rosiglitazone as groups according to the combination of metformin and OHA: antidiabetic monotherapies in drug-naïve, Korean T2D pa- group I (metformin and either gliclazide-MR or glimepiride), tients. The study showed no statistical difference in the efficacy group II (metformin and 15 mg pioglitazone), and group III of glimepiride, metformin, and rosiglitazone as antidiabetic (metformin and 100 mg sitagliptin).
monotherapy. However, there has been no report on the effi- Complete available medical records from other departments cacy or safety of metformin-based dual combination therapy or other institutions were reviewed, and laboratory results for drug-naïve or newly detected Korean T2D patients.
meeting the criteria for diabetes (fasting serum glucose ≥126 This study was designed to evaluate the effectiveness of gly- mg/dL, postprandial serum glucose ≥200 mg/dL, or HbA1c cemic control in drug-naïve or newly detected Korean T2D ≥6.5%) were regarded as the onset of diabetes. If data were patients receiving metformin-based dual combination therapy unavailable, the onset and duration of diabetes were deter-with SU, pioglitazone, or DPP4-inhibitor.
mined by subject recal . Subjects underwent a standardized mixed-meal stimulation test (Glucerna; Abbott Laboratories, Saint-Laurent, QC, Canada) (H4S 1Z; 2 cans, total 474 mL, 474 kcal, 26 g fat, 45.8 g carbohydrate, and 19.8 g protein) for Study design
the evaluation of glucose metabolism. The glucose level after a This prospective, nonrandomized, open-label study was con- mixed-meal stimulation test was regarded as the baseline post- ducted at a single center by closely observing metabolic pa- prandial glucose (PPG) level. Plasma glucose level was mea- rameters for up to 24 weeks between November 2011 and sured using the glucose oxidase method, and HbA1c was mea-March 2013. The study protocol entitled ‘Efficacy of antidia- sured with high-performance liquid chromatography using betic medications recommended by government guidelines Variant II Turbo (Bio-Rad Laboratories, Hercules, CA, USA). for newly diagnosed or currently medicated T2D patients on Plasma triglyceride, total cholesterol, high density lipoprotein metformin and sulfonylurea’ was reviewed by the local ethics cholesterol, blood urea nitrogen, creatinine, AST, and ALT lev-committee (2011-0670-001). To adhere to the guidelines of els were assayed with a routine Hitachi 7600 autoanalyzer (Hit-NHIC and Institutional Review Board for reimbursement, all achi Instruments Service, Tokyo, Japan). Low density lipopro-subjects with an initial HbA1c level ≥7.5% received metformin tein cholesterol level was calculated using Friedewald’s equa-and were recommended lifestyle modification; however, the tion. Serum insulin and C-peptide levels were measured in du-selection of another OHA from SU, pioglitazone, or DPP4-in- plicate using immunoradiometric assays (Beckman Coulter, hibitor was at the discretion of the subjects’ physicians. In ad- Fullerton, CA, USA) with samples individual y collected during dition, physicians also determined all subsequent treatment the standardized mixed-meal stimulation test. Pancreatic β-cell Response to metformin-based dual combinations function and insulin sensitivity were determined by homeosta- (≤7.0%) level according to baseline HbA1c.
sis model assessment (HOMA) of pancreatic β-cell function (HOMA-β; [baseline insulin (μIU/mL)×20]/[0.055551×fast- Statistics
ing glucose (mg/dL)–3.5]) and HOMA of insulin resistance Results are described as mean±standard deviation or median (HOMA-IR; [0.055551×fasting glucose (mg/L)×baseline insu- value (low quartile, high quartile). Analysis of variance test was used for comparison of baseline data among the three treatment groups. Kruskal-Wallis test was used for nonpara- Tolerability and hypoglycemia assessment
metric statistical analysis. Mann-Whitney test with Bonferroni Reasons for discontinuation of the study and treatment-in- correction was used as post hoc analysis for nonparametric duced major hypoglycemia were recorded to assess tolerability statistical analysis. Wilcoxon signed rank test was used for and compliance. A major hypoglycemic event was defined as comparison of pretreatment and posttreatment values. Multi-blood glucose ≤60 mg/dL accompanied by neurological symp- ple linear regression test and Fisher exact test were performed toms consistent with hypoglycemia or an episode requiring for comparison of treatment effectiveness. Group I was used intervention with intravenous glucose. In addition, other mi- as a reference group in the multiple regression test because nor adverse events were obtained by patient self-report.
dual therapy with sulfonylurea and metformin is the most tra-ditional combination. Statistical analyses were performed us- Effectiveness assessment
ing SPSS version 20 (IBM Co., Armonk, NY, USA). Differenc- The primary endpoint was change in HbA1c and fasting and es among groups with P<0.05 were considered statistical y PPG levels from baseline to 24 weeks. The secondary endpoint significant.
was the frequency of successfully achieved target HbA1c Fig. 1. Flow of the study. A total of 116 patients were enrolled, and 99 subjects were analyzed in the study. Of these patients, 28 in
the glimepiride/metformin group, 27 in the pioglitazone/metformin group, and 33 in the sitagliptin/metformin group complet- ed the study without medication change. GI, gastrointestinal.
60) mg] or glimepiride n=9 [29.0%], 4 [2.5 to 4] mg); group II (n=30, metformin and 15 mg pioglitazone), and group III Baseline characteristics of the study population
(n=38, metformin and 100 mg sitagliptin) (Fig. 1). The base- Of the 116 subjects who were enrolled in this study, 17 subjects line characteristics of the subjects are described in Table 1. The failed to complete follow-up; 99 subjects were ultimately ana- mean age, body mass index (BMI), and median HbA1c level lyzed and were classified into three groups: group I (n=31, of the study population were 53.2 years, 26.7 kg/m2, and 9.1%, metformin and either gliclazide-MR [n=22 (71.0%), 60 (30 to respectively. No significant difference in diabetes duration or Table 1. Baseline characteristics and daily metformin dose
Values are presented as mean±standard deviation, number (%), or median (low quartile, high quartile). Analysis of variance test was used for parametric analysis and Kruskal-Wallis test for nonparametric analysis.
BMI, body mass index; FBS, fasting blood glucose; PPG, postprandial glucose; HbA1c, hemoglobin A1c; HDL-C, high density lipoprotein cho- lesterol; LDL-C, low density lipoprotein cholesterol; HOMA-β, homeostasis model assessment of β-cell function; HOMA-IR, homeostasis model assessment of insulin resistance; BUN, blood urea nitrogen; AST, aspartate aminotransferase; ALT, alanine aminotransferase.
Response to metformin-based dual combinations baseline demographics, anthropometrics, or metabolic char- acteristics was observed among the three groups. Daily met- formin dose was smaller in group I than the other two groups (overal , P=0.006; group I vs. group II, P=0.001; group I vs. group III, P=0.012; group II vs. group III, P=0.064).
Tolerability and hypoglycemia assessment
The proportion of participants who completed the study with- out change in dose or class of initial medications was 90.3% in the SU-treated group I (28/31), 90% in the pioglitazone-treat- ed group II (27/30), and 86.8% in the sitagliptin-treated group III (33/38). There was no significant difference in study com- Fig. 2. Change in hemoglobin A1c (HbA1c) from baseline to
pletion rate among the groups (P=0.925). In group I, one sub- 24 weeks. Solid lines indicate the median HbA1c level, and ject discontinued the study due to very good response and broken lines indicate the median of individual y assessed dif- ferences in HbA1c level from baseline. Multiple linear regres- subsequent reduction of medications and another due to in- sion was used for statistical analysis, and group I was used as a sufficient response. In addition, one subject with symptoms of reference group. The change of HbA1c was not statistical y dif- hypoglycemia was prescribed other drugs. In group II, one ferent among groups at 12 weeks (group II, P=0.101; group III, subject with an insufficient response, one subject with nausea, P=0.673) or 24 weeks (group II, P=0.066; group III, P=0.678) and one subject concerned with the risk of bladder cancer after adjustment for baseline age, sex, body mass index, and were switched to other drugs. In group III, one subject with a baseline HbA1c level.
very good response, three subjects with an insufficient response, and one subject with nausea were switched to other drugs. No week (group II, P=0.101; group III, P=0.673) or 24-week end major hypoglycemic events occurred among the groups. In point (group II, P=0.066; group III, P=0.678) (Table 2, Fig. 2). addition, one patient in group II reported transient diarrhea, However, after additional adjustment of metformin dose, group but no other adverse events were reported.
I showed superiority in HbA1c improvement to group II at the 24-week end point (covariate-adjusted difference in change of Effectiveness assessment
HbA1c, 0.35%; P=0.046) (Table 2). Fasting plasma glucose Primary outcome
level decreased from 166.5 (139.0 to 195.0) to 103.5 mg/dL In the 88 patients who completed the study without change of (89.0 to 112.0) (P<0.001); 174.0 (145.0 to 223.0) to 111.0 mg/dL initial medications, the median HbA1c level decreased from (101.5 to 120.0) (P<0.001); and 173.0 (135.0 to 204.0) to 105.0 8.9% (8.2 to 10.3) to 6.5% (6.4 to 7.0; at 12-week) and to 6.4% mg/dL (100.0 to 124.0) (P<0.001) in groups I, II, and III, re-(6.0 to 6.7; at 24-week) in the SU-treated group I (P<0.001 for spectively. PPG level decreased from 226.5 (192.5 to 312.0) to each); from 9.0% (8.4 to 11.2) to 6.8% (6.5 to 7.3; at 12-week) 157.0 mg/dL (133.5 to 196.5) (P<0.001); 238.0 (195.5 to 324.0) and to 6.6% (6.1 to 6.9; at 24 weeks) in the pioglitazone-treated to 157.0 mg/dL (124.0 to 219.5) (P<0.001); and 251.0 (196.0 group II (P<0.001 for each); and from 9.3% (7.8 to 10.4) to 6.4% to 306.0) to 148.0 mg/dL (115.0 to 172.0) (P<0.001) in groups (6.3 to 7.0; at 12-week) and to 6.3% (6.0 to 6.7; at 24-week) (P< I, II, and III, respectively. In addition, there was no significant 0.001 for each) in the sitagliptin-treated group III (Fig. 2). The difference in the change of FBG (group II, P=0.061; group III, median of individual y assessed differences in HbA1c level P=0.070) or PPG (group II, P=0.914; group III, P=0.237) lev-from baseline to the 12- and 24-week end points were -2.2% el among the groups after adjusting for baseline age, sex, BMI, (-3.7 to -1.8) and -2.5% (-4.0 to -1.9) in group I; -2.2% (-3.8 to and FBG or PPG. After additional adjustment of metformin -1.5) and -2.8% (-4.5 to -1.6) in group II; and -2.1% (-4.0 to dose, group I showed superiority in FBG improvement to -1.6) and -2.7% (-4.0 to -1.6) in group III, respectively. In ad- group II and group III (covariate-adjusted difference in change dition, there was no significant difference in the change of of FBG: group 2, 15.9 mg/dL, P=0.008; group 3, 11.6 mg/dL, HbA1c level of group II and III compared to group I after ad- P=0.032), without significant difference in the change of PPG justing for baseline age, sex, BMI, and HbA1c at either the 12- Table 2. Covariate-adjusted differences in change of hemoglo-
Proportions of individuals achieving HbA1c ≤7% Group II (n=27) Group III (n=33) Proportions of overall successful diabetes control 0 Group I (n=31) Group II (n=30) Group III (n=38) B
Fig. 3. (A) The proportions of individuals who achieved hemo-
globin A1c (HbA1c) level ≤7% at 24 weeks, and (B) the pro- portions of individuals with overall successful diabetes control according to medication. (A) Dark gray bar indicates the pro- portion of individuals who achieved an HbA1c level ≤7% among those who did not change medications by the end of the study. Fisher exact test was used to compare the three groups, and there was no significant difference (P=0.649). (B) Light gray bar indicates the proportion of individuals, among the to- tal number of subjects who initial y enrolled in the study, except In model 1, which adjusts for baseline age, sex, BMI, and HbA1c, there for those who did not continue with follow-up observation, was no significant difference between groups in change of HbA1c who achieved an HbA1c level ≤7% at 24 weeks or reduced (group II, P=0.066; group III, P=0.678). Additional adjustment of HOMA-β and HOMA-IR did not alter the statistical significance. medication during the study period because of a very good re- However, after additional adjustment of daily metformin dose, group sponse. The Fisher exact test was used to compare the three II showed a statistical y smaller reduction in HbA1c than group I groups, and there was no significant difference (P=0.593).
(covariate-adjusted difference in change of HbA1c, 0.35%; P=0.046). HbA1c, hemoglobin A1c; BMI, body mass index; HOMA-β, homeo- who reduced their medications because of a very good response stasis model assessment of β-cell function; HOMA-IR, homeostasis model assessment of insulin resistance.
during the study period were considered to have successful y controlled diabetes; however, other subjects, with the excep- The proportion of individuals who achieved HbA1c level tion of those who did not continue with follow-up observation, ≤7% at 24 weeks was 89.3%, 81.5%, and 84.8% in groups I, II, were considered to have failed diabetes control. The proportion and III, respectively, with no statistical y significant difference of individuals with successful y controlled diabetes was 83.9%, in the proportion among the groups (P=0.649) (Fig. 3A). To 73.3%, and 76.3% in groups I, II, and III, respectively, with no minimize selection bias, we performed an additional analysis. statistical y significant difference among the groups (P=0.593) Subjects who achieved HbA1c level ≤7% at 24 weeks and those (Fig. 3B).
Response to metformin-based dual combinations Table 3. Baseline characteristics and daily metformin dose of three categories according to baseline hemoglobin A1c level
Values are presented as median (low quartile, high quartile), mean±standard deviation, or number (%). Analysis of variance test was used for parametric analysis and Kruskal-Wallis test for nonparametric analysis.
BMI, body mass index; HbA1c, hemoglobin A1c; HOMA-β, homeostasis model assessment of β-cell function; HOMA-IR, homeostasis model aStatistical significance (P<0.017) in post hoc analysis (Mann-Whitney test with Bonferroni correction) was indicated by category I vs. III, bCat- egory II vs. III, cCategory I vs. II.
Secondary and other outcomes
The subjects were also classified into three categories accord- ing to HbA1c level: category I (n=43; 7.5%≤HbA1c<9.0%), category II (n=23; 9.0%≤HbA1c<11.0), and category III (n=22; 11.0%≤HbA1c) (Table 3). In addition to different HbA1c level (8.1% vs. 9.9% vs. 11.6%; P<0.001), HOMA- (34.2 vs. 25.6 vs. 19.8; P=0.008), and HOMA-IR (2.8 vs. 3.6 vs. 3.5; P=0.005) were also significantly different among the cate- gories. The daily dose of metformin was higher in category III 0 Category I (n=43) Category II (n=23) Category III (n=22) than in the other two categories (1,000 mg vs. 1,000 mg vs.
1,700 mg; P<0.001). Despite the difference in baseline HbA1c Fig. 4. Change in hemoglobin A1c (HbA1c) according to base-
level and HOMA, no statistical y significant difference was ob-
line HbA1c level. Dark gray bar indicates the level at baseline and light gray bar at 24 weeks. Bars represent medians (low served among the categories after 24 weeks (6.4% [6.1 to 6.8] quartile, high quartile). Despite different initial HbA1c levels vs. 6.6% [6.2 to 6.7] vs. 6.0% [5.7 to 6.6)], P=0.051) (Fig. 4). (8.2% vs. 9.9% vs. 11.9%; P<0.001), HbA1c level in each cate- The proportion of individuals who achieved HbA1c ≤7% at gory after 24 weeks of treatment demonstrated no statistical 24 weeks was 83.7%, 87.0%, and 86.4% in categories I, II, and difference by Kruskal-Wallis test (6.4% vs. 6.6% vs. 6.0%; P= III, respectively, with no statistical y significant difference 0.051).
among the categories (P=1.000) (Fig. 5A). The proportion of individuals with overall successful diabetes control, as defined OHA therapy and its glucose-lowering efficacy have been es-above, was 75.0%, 81.5%, 79.2% and similar among categories pecial y lacking in Korean subjects with T2D. On the basis of I, II, and III, respectively (P=0.819) (Fig. 5B).
previous research, this study focused on the practical implica-tions of a combination regimen for metformin-based OHA DISCUSSION
therapy in newly diagnosed or drug-naïve Korean T2D pa-tients. Therefore, we attempted to investigate: 1) the tolerabili- It is well known that significant prevention and reduction of mi- ty and efficacy of metformin-based dual combination therapy crovascular and macrovascular complications can be achieved with OHAs by assessing the reduction of HbA1c level as well with early intensive glycemic control in subjects with newly as the proportion of subjects reaching a target HbA1c level detected T2D [1,3]. However, to date, scientific studies investi- ≤7%; and 2) the appropriateness of current guidelines estab- gating an optimal combination regimen for metformin-based lished by the NHIC of Korea, which mandates the number of Proportions of individuals achieving HbA1c ≤7% Proportions of overall successful diabetes control Category I (n=43) Category II (n=23) Category III (n=22) A
Category I (n=48) Category II (n=27) Category III (n=24) B
Fig. 5. (A) The proportions of individuals who achieved an hemoglobin A1c (HbA1c) level ≤7% at 24 weeks, and (B) the propor-
tions of individuals with overall successful diabetes control according to baseline HbA1c level. Category I, 7.5%≤HbA1c<9.0; category II, 9.0%≤HbA1c<11.0; category III, 11.0%≤ HbA1c. (A) Dark gray bar indicates the proportion of individuals who achieved an HbA1c level ≤7% among those who did not change medications by the end of the study. Fisher exact test was used to compare the three groups, and there was no significant difference (P=1.000). (B) Light gray bar indicates the proportion of in- dividuals, among the total number subjects who initial y enrolled in the study, except for those who did not continue with follow- up observation, who achieved an HbA1c level ≤7% at 24 weeks or reduced medication during the study period because of a very good response. Fisher exact test was used to compare the three groups, and there was no significant difference (P=0.819).
OHAs prescribed by analyzing the proportion of subjects two drugs, which is similar to our results [10-12]. The HOMA-reaching a target HbA1c level ≤7% according to initial HbA1c.
IR of our study (3.2) was lower than those of previous studies In this study, 99 Korean T2D subjects who were newly diag- (3.6 to 7.0) [12-14,16]. Even though HOMA-IR has limited nosed with the disease or who were drug-naïve underwent value when compared across different populations [17], this dual therapy with a combination of metformin and OHA (SU finding may suggest that our subjects had better insulin sensi-[either gliclazide-MR or glimepiride], 15 mg pioglitazone, or tivity than subjects in previous studies.
100 mg sitagliptin) over a 24-week study period. By the end of In the aspects of tolerability and glycemic effectiveness, the the study, the percentage of subject with initial HbA1c level three dual therapies using a combination of metformin and ei- ≥7.5% decreased by 2.5% to 2.7%. In previous studies, most ther sulfonylurea, pioglitazone, or sitagliptin led to similar OHAs decreased HbA1c level by about 1% when used as a proportions of subjects adhering to the initial regimen, similar monotherapy, and by about 2% when used as a combination degrees of HbA1c improvement, and similar proportions of therapy [8,9]. The differences between our results and those of subjects reaching the target HbA1c level. However, after ad-previous studies might be due to the higher baseline HbA1c justment of daily metformin dose, which was determined ac-levels, drug-naïve characteristics, and lower insulin resistance cording to physician judgment in usual practice, sulfonylurea of our subjects. Our study did not exclude subjects with very reduced HbA1c by a greater magnitude than did pioglitazone. high HbA1c level, unlike many previous studies which set an This finding suggests that pioglitazone and metformin combi-upper limit of initial HbA1c [10-12]. Moreover, the baseline nation could partial y overcome the lower intrinsic potency of HbA1c (9.1%) of our study was higher than those of most pre- pioglitazone by using a higher dose of metformin in real prac- vious studies (7.6% to 8.8%) [10-15]. Because initial HbA1c tice.
level did not alter the endpoint HbA1c level in our study, this Interestingly, baseline HbA1c level did not affect the endpoint higher baseline HbA1c seems to have mostly affected the de- HbA1c level in this study. Moreover, the proportion of indi- gree of HbA1c reduction. Contrary to many studies evaluating viduals who reached a target HbA1c level ≤7.0% was similar patients taking OHA prior to the study, which reported higher across the groups, although their baseline HbA1c levels were HbA1c (6.9% to 7.2%) at endpoint than our study [13,14], sev- significantly different. It is well known that patients with high- eral studies evaluating drug-naïve diabetes patients observed er baseline HbA1c level have greater reduction of HbA1c irre-final HbA1c levels of about 6.5% after using a combination of spective of drug class [18,19]. Because each drug results in a Response to metformin-based dual combinations greater reduction of HbA1c according to higher baseline HbA1c, tial HbA1c level ≥7.5%. This study is the first to compare the when two drugs are used simultaneously, the increased reduc- glycemic effectiveness of dual combination agents commonly tion of HbA1c according to higher baseline HbA1c might used as primary medications in Korean T2D patients. An ad-show a greater magnitude due to the additive effect. This ten- ditional randomized study with a larger number of subjects is dency was also shown in previous studies, even though statis- warranted to obtain more detailed information including glu- tical analysis was not performed [10,19]. Another possible ex- planation is preserved insulin sensitivity of the study patients.
HOMA-IR of subjects with the highest initial HbA1c level was CONFLICTS OF INTEREST
3.5 in this study, which was relatively lower than reported in
other studies (3.6 to 7.0) which have assessed the efficacy of No potential conflict of interest relevant to this article was re-
dual regimens [12-15]. In our study, a one unit increase of ported.
HOMA-IR resulted in a 0.066%, increase of HbA1c at end-
point. The other explanation is the higher metformin dose for REFERENCES
patients with higher initial HbA1c level. These results indicate that, even at a very high baseline HbA1c level, therapy with a 1. UK Prospective Diabetes Study (UKPDS) Group. Effect of in- combination of two drugs may still be effective in drug-naïve tensive blood-glucose control with metformin on complications subjects whose insulin sensitivity remains preserved. This in overweight patients with type 2 diabetes (UKPDS 34). Lan- finding supports the current guidelines for OHA selection, which do not recommend initial therapy using a combination 2. Holman RR, Paul SK, Bethel MA, Neil HA, Matthews DR. Long-term follow-up after tight control of blood pressure in This study has several limitations. The patients were not type 2 diabetes. N Engl J Med 2008;359:1565-76.
randomized, and the criteria for changing medications were 3. Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, Motoyo- not uniform because all treatment decisions were determined shi S, Kojima Y, Furuyoshi N, Shichiri M. Intensive insulin by diabetologists specifical y for each patient. However, the therapy prevents the progression of diabetic microvascular baseline characteristics of the three groups did not show statis- complications in Japanese patients with non-insulin-depen- tical y significant differences. Furthermore, supplemental anal- dent diabetes mellitus: a randomized prospective 6-year study. ysis of all subjects, except for those who did not continue with Diabetes Res Clin Pract 1995;28:103-17.
follow-up observation, showed similar statistical results to the 4. Rodbard HW, Jel inger PS, Davidson JA, Einhorn D, Garber analysis of subjects who did not change their medications AJ, Grunberger G, Handelsman Y, Horton ES, Lebovitz H, throughout the study. In addition, the number of participants Levy P, Moghissi ES, Schwartz SS. Statement by an American was smal , and the rate of follow-up loss was 14.7% at the end Association of Clinical Endocrinologists/American College of of the study. Therefore, the reliability of our trial was lower Endocrinology consensus panel on type 2 diabetes mellitus: an than original y expected at the time the study was first de- algorithm for glycemic control. Endocr Pract 2009;15:540-59.
5. Ko SH, Kim SR, Kim DJ, Oh SJ, Lee HJ, Shim KH, Woo MH, In conclusion, metformin-based dual combination thera- Kim JY, Kim NH, Kim JT, Kim CH, Kim HJ, Jeong IK, Hong pies with OHAs including sulfonylurea, pioglitazone, or sita- EK, Cho JH, Mok JO, Yoon KH; Commit ee of Clinical Prac- gliptin showed similar glycemic effectiveness in drug-naïve tice Guidelines, Korean Diabetes Association. 2011 clinical Korean subjects with newly diagnosed T2D. Combination practice guidelines for type 2 diabetes in Korea. Diabetes therapy using these OHA drugs was similarly effective in pa- tients with a wide range of initial HbA1c level. Based on these 6. Inzucchi SE, Bergenstal RM, Buse JB, Diamant M, Ferrannini results, we suggest the appropriateness of the current guide- E, Nauck M, Peters AL, Tsapas A, Wender R, Matthews DR; lines established by the NHIC of Korea, which do not allow American Diabetes Association (ADA); European Association initial three drug combinations and recommend metformin- for the Study of Diabetes (EASD). Management of hyperglyce- based dual combination therapy with OHAs including sulfo- mia in type 2 diabetes: a patient-centered approach: position nylurea, pioglitazone, or DPP4-inhibitor in subjects with ini- statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). tial combination therapy with sitagliptin, a dipeptidyl pepti- dase-4 inhibitor, and metformin on glycemic control in pa- 7. Yoon KH, Shin JA, Kwon HS, Lee SH, Min KW, Ahn YB, Yoo tients with type 2 diabetes. Diabetes Care 2007;30:1979-87.
SJ, Ahn KJ, Park SW, Lee KW, Sung YA, Park TS, Kim MS, Kim 13. Tosi F, Muggeo M, Brun E, Spiazzi G, Perobelli L, Zanolin E, YK, Nam MS, Kim HS, Park Ie B, Park JS, Woo JT, Son HY. Gori M, Coppini A, Moghetti P. Combination treatment with Comparison of the efficacy of glimepiride, metformin, and metformin and glibenclamide versus single-drug therapies in rosiglitazone monotherapy in Korean drug-naive type 2 dia- type 2 diabetes mellitus: a randomized, double-blind, compar- betic patients: the practical evidence of antidiabetic mono- ative study. Metabolism 2003;52:862-7.
therapy study. Diabetes Metab J 2011;35:26-33.
14. Kaku K. Efficacy and safety of therapy with metformin plus pi- 8. Bennett WL, Maruthur NM, Singh S, Segal JB, Wilson LM, oglitazone in the treatment of patients with type 2 diabetes: a Chatterjee R, Marinopoulos SS, Puhan MA, Ranasinghe P, double-blind, placebo-controlled, clinical trial. Curr Med Res Block L, Nicholson WK, Hutfless S, Bass EB, Bolen S. Com- parative effectiveness and safety of medications for type 2 dia- 15. Rhee EJ, Lee WY, Min KW, Shivane VK, Sosale AR, Jang HC, betes: an update including new drugs and 2-drug combinations. Chung CH, Nam-Goong IS, Kim JA, Kim SW; Gemigliptin Study 006 Group. Efficacy and safety of the dipeptidyl pepti- 9. Bennett WL, Balfe LM, Faysal JM. AHRQ’s comparative effec- dase-4 inhibitor gemigliptin compared with sitagliptin added tiveness research on oral medications for type 2 diabetes: a to ongoing metformin therapy in patients with type 2 diabetes summary of the key findings. J Manag Care Pharm 2012;18(1 inadequately controlled with metformin alone. Diabetes Obes 10. Garber AJ, Donovan DS Jr, Dandona P, Bruce S, Park JS. Effi- 16. Rhee SY, Woo JT. The prediabetic period: review of clinical as- cacy of glyburide/metformin tablets compared with initial pects. Diabetes Metab J 2011;35:107-16.
monotherapy in type 2 diabetes. J Clin Endocrinol Metab 2003; 17. Wal ace TM, Levy JC, Mat hews DR. Use and abuse of HOMA 88:3598-604.
modeling. Diabetes Care 2004;27:1487-95.
11. Rosenstock J, Rood J, Cobitz A, Biswas N, Chou H, Garber A. 18. Bloomgarden ZT, Dodis R, Viscoli CM, Holmboe ES, Inzucchi Initial treatment with rosiglitazone/metformin fixed-dose SE. Lower baseline glycemia reduces apparent oral agent glucose- combination therapy compared with monotherapy with either lowering efficacy: a meta-regression analysis. Diabetes Care rosiglitazone or metformin in patients with uncontrolled type 2 diabetes. Diabetes Obes Metab 2006;8:650-60.
19. Deacon CF. Dipeptidyl peptidase-4 inhibitors in the treatment 12. Goldstein BJ, Feinglos MN, Lunceford JK, Johnson J, Wil- of type 2 diabetes: a comparative review. Diabetes Obes Metab liams-Herman DE; Sitagliptin 036 Study Group. Effect of ini-

Source: http://www.e-dmj.org/Synapse/Data/PDFData/2004DMJ/dmj-37-465.pdf