Clinical Care/Education/Nutrition ORIGINAL ARTICLE

The Synergistic Effect of Miglitol Plus Metformin Combination Therapy in the Treatment of Type 2 Diabetes

1 JEAN-LOUIS CHIASSON, MD The recently developed class of 2 LISA NADITCH, MD ␣-glucosidase inhibitors has a unique FOR THE MIGLITOL CANADIAN UNIVERSITY mode of action; it blocks oligosaccharide INVESTIGATOR GROUP catabolism, delays carbohydrate diges- tion and absorption, and smooths and lowers postprandial plasma blood glu- cose peaks (6–9). Substantial evidence supports their use as monotherapy or ad- OBJECTIVE — To investigate the efficacy and safety of miglitol in combination with met- junct therapy for poorly controlled type 2 formin in improving glycemic control in outpatients in whom type 2 diabetes is insufficiently diabetes (5,10–14). Miglitol is the first controlled by diet alone. pseudomonosaccharide ␣-glucosidase in- RESEARCH DESIGN AND METHODS — In this multicenter, double-blind, placebo- hibitor derived from 1-deoxynojirimycin controlled study, 324 patients with type 2 diabetes were randomized, after an 8-week placebo and is structurally a glucose analog run-in period, to treatment with either placebo, miglitol alone, metformin alone, or miglitol plus (15,16). Its efficacy in monotherapy metformin for 36 weeks. The miglitol was titrated to 100 mg three times a day and metformin was (13,17) and in combination with sulfo- administered at 500 mg three times a day. The primary efficacy criterion was change in HbA1c nylureas (11) as a glucose-lowering agent from baseline to the end of treatment. Secondary parameters included changes in fasting and in type 2 diabetes has been shown in a postprandial plasma glucose and insulin levels, serum triglyceride levels, and responder rate. number of clinical studies. A study of miglitol in combination with metformin RESULTS — A total of 318 patients were valid for intent-to-treat analysis. A reduction in mean Ϫ has been reported previously in elderly placebo-subtracted HbA1c of 1.78% was observed with miglitol plus metformin combination therapy, which was significantly different from treatment with metformin alone (Ϫ1.25; P ϭ type 2 diabetic patients (18). However, it 0.002). Miglitol plus metformin also resulted in better metabolic control than metformin alone is still unclear whether miglitol can en- for fasting plasma glucose (Ϫ44.8 vs. Ϫ20.4 mg/dl; P ϭ 0.0025), 2-h postprandial glucose area hance glycemic control when given in under the curve (Ϫ59.0 vs. –18.0 mg/dl; P ϭ 0.0001), and responder rate (70.6 vs. 45.52%; P ϭ combination with metformin in middle- 0.0014). All therapies were well tolerated. aged type 2 diabetic patients and whether the safety and tolerability profile of migli- CONCLUSIONS — In type 2 diabetic patients, miglitol in combination with metformin tol and metformin as monotherapy is af- gives greater glycemic improvement than metformin monotherapy. fected by such a combination. Therefore, this study was performed to investigate Diabetes Care 24:989–994, 2001 the efficacy and safety of miglitol in com- bination with metformin in improving glycemic control, compared with met- aintaining normal plasma glucose in combination. The biguanide met- formin monotherapy, in middle-aged levels is a key factor in reducing formin is a frequent first-line choice of outpatients in whom type 2 diabetes was M the risk of developing diabetes antidiabetic medication (3,4). However, insufficiently controlled by diet alone. complications (1,2). Current recommen- monotherapy with any hypoglycemic The primary control group in this study dations supported by recent data from the agent eventually necessitates the use of in- was the metformin arm. Comparisons U.K. Prospective Diabetes Study data em- creasing doses because type 2 diabetes with miglitol monotherapy and placebo phasize life-style management, diet, and worsens over time with declining pancre- were also performed. exercise as the first-line approach, fol- atic ␤-cell function (5) and eventually re- lowed by therapy with hypoglycemic or quires addition of a second antidiabetic RESEARCH DESIGN AND antihyperglycemic agents, either alone or medication. METHODS — This was a multicenter, ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● double-blind, randomized, placebo- From the 1Research Group on Diabetes and Metabolic Regulation, Research Center, Centre Hospitalier de 2 controlled parallel group study. Eligible l’Universite´ de Montre´al, Montreal, Quebec, Canada; and Sanofi-Synthelabo, Paris, France. patients were type 2 diabetic patients Address correspondence and reprint requests to Jean-Louis Chiasson, MD, Research Group on Diabetes Ͼ and Metabolic Regulation, Research Center CHUM, Hoˆtel-Dieu CHUM, 3850 St. Urbain, 8–202, Montreal, (men and women) 40 years of age Quebec, Canada H2W 1T8. E-mail: [email protected]. in whom diabetes was inadequately Received for publication 28 November 2000 and accepted in revised form 6 February 2001. controlled by diet alone, i.e., in whom L.N. is an employee of Sanofi-Synthelabo, which is involved in the marketing of the product miglitol for HbA level was Ն7.2 and Յ9.5%. The the treatment of type 2 diabetes. 1c Abbreviations: AUC, area under the curve; ITT, intent to treat. major exclusion criteria included type 1 A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion diabetes, the presence of major debili- factors for many substances. tating diseases, recent cardiovascular

DIABETES CARE, VOLUME 24, NUMBER 6, JUNE 2001 989 Miglitol plus metformin in type 2 diabetes events, gastrointestinal diseases, medica- tions were given with the tolerance test tory, vital signs, and concomitant medi- tions likely to affect intestinal motility or meal. The proportion of responders in the cations. The primary efficacy parameter the absorption of nutrients, hypersensi- different treatment groups was also eval- was the change in HbA1c from baseline to tivity to miglitol or metformin, and a his- uated, in which a clinically significant re- the end of treatment and was analyzed by tory of lactic acidosis. Patients who were sponse was defined as either a Ն15% analysis of variance techniques. All four taking either sulfonylurea or metformin reduction in HbA1c from baseline or an treatment groups were included in the Ͻ below the maximum dose could be eligi- HbA1c level 7.0%. primary comparisons, which used or- ble as long as the antidiabetic drug was HbA1c was measured during screen- thogonal contrasts. discontinued. Before starting any study ing, at week –2 and baseline, and there- The primary analysis of efficacy was procedures, patients who fulfilled the el- after at each visit (weeks 4, 8, 12, 16, performed on the ITT population, which igibility criteria gave their written in- 20, 28, and 36) after randomization. Rou- included all patients who had both base- formed consent and entered a single- tine fasting plasma glucose was mea- line and postrandomization efficacy val- blind, 8-week placebo run-in period. All sured at week Ϫ2 and at every subsequent ues. For end-of-treatment analysis, the subjects were seen by a dietitian before visit. Fasting and postprandial (60, 90, last-observation-carried-forward method the run-in period and were advised re- and 120 min) glucose and insulin lev- was used. All patients who received at garding a well-balanced weight-reducing els as well as triglyceride levels were least one dose of trial medication were in- diet (19). They were also advised regard- specifically measured at baseline and at cluded in the safety analysis. Adverse ing exercise, mainly walking 20–30 min weeks 12, 16, and 36 or at premature events and other safety parameters were at least three times per week. The diet was termination. analyzed in terms of the percentage of pa- reinforced after 1 month, and if the HbA1c Safety and tolerability were investi- tients in whom they occurred. was still Ն7.2% by 2 months, they were gated by the occurrence of adverse events; considered dietary failure. Patients with hypoglycemic events; changes in vital RESULTS — A total of 324 patients Ն Յ HbA1c levels 7.2 but 9.5% (110– signs; and changes in laboratory values, were randomized in the study: 83 to pla- 146% above the upper limit of the normal including biochemical parameters, hema- cebo, 82 to miglitol monotherapy, 83 to range) at the end of the run-in period tology, vitamins (thiamine-dependent metformin monotherapy, and 76 to met- were eligible for randomization to match- transketolase, folate, vitamin B12, vitamin formin/miglitol combination therapy. Six ing placebo or to active treatment with A, and retinol protein binding), and stan- patients (one each from the placebo and either miglitol plus placebo, metformin dard urinalysis. Adverse events were metformin/miglitol combination therapy plus placebo, or metformin plus miglitol monitored and documented at week –2 groups and two each from the miglitol for 36 weeks. The miglitol dosage was and at all visits thereafter. Routine bio- and metformin monotherapy groups) force-titrated as follows: administration chemistry, hematology, and urinalysis were excluded from the ITT analysis be- of the drug was started at 25 mg three were performed at screening, baseline, cause of the absence of recorded HbA1c times a day for 4 weeks, increased to 50 and weeks 12, 20, and 36 or at premature values after randomization. A total of 318 mg three times a day for 8 weeks, and termination; vitamin assays were per- patients (98%) were valid for the ITT then increased to 100 mg three times a formed at baseline and at the final visit. analysis. Selected demographic parame- day until the end of the study. Adminis- Physical examinations were conducted at ters and clinical data for all randomized tration of metformin was started and screening, week –2, baseline, week 12, patients are shown in Table 1. Treatment maintained at 500 mg three times a day and the end of the study, whereas vi- groups were comparable for number of throughout the study. All patients were tal signs and weight were recorded at each patients, age, race (predominately Cauca- given an Elite Glucometer and were ad- visit. Routine laboratory determinations sian), weight and BMI, sex ratio, duration vised regarding regular home blood- were performed using standard method- of diabetes, and baseline HbA1c. The glucose monitoring. ology by local laboratories, whereas the mean durations of treatment exposure by preprandial and postprandial efficacy treatment group were as follows: placebo Efficacy and safety evaluation measurements and vitamin evaluations 200.3 Ϯ 93.3 days; miglitol 190.9 Ϯ 90.5 The primary efficacy criterion was the were performed at a central laboratory. days; metformin 231.6 Ϯ 70.1 days; met- change in HbA1c level from baseline to the formin plus miglitol combination end of the double-blind treatment for the Statistical analysis 203.3 Ϯ 94.6 days. intent-to-treat (ITT) population, which To detect a mean difference of 0.6% in the The primary comparison for all effi- included any patient who had both a change in HbA1c between treatment cacy parameters was between the met- baseline value and at least one postran- groups from baseline to the end of treat- formin plus miglitol combination and domization efficacy value. ment, with ␣ϭ0.05 and a power (1–␤)of metformin monotherapy treatment Secondary efficacy parameters in- at least 90%, a minimum of 60 patients groups. The mean changes in HbA1c val- cluded the change in fasting and post- per treatment group were required. To ac- ues (means Ϯ SEM) in response to the prandial plasma glucose and insulin levels count for multicenter variability, this was different treatments are shown in Table 2. and serum triglyceride levels from base- increased by 20% to 75 patients per treat- There was an increase of 0.38 Ϯ 0.12% in line to the end of treatment, measured at ment arm. the placebo group, virtually no change in 0, 60, 90, and 120 min after a standard- Treatment groups were compared at the miglitol group (0.02 Ϯ 0.10%), a de- ized liquid test breakfast (55% carbohy- baseline for demographic variables and crease of Ϫ0.85 Ϯ 0.12% in the met- drate, 30% fat, and 15% protein; other prognostic factors, such as family formin group, and a decrease of Ϫ1.39 Ϯ providing ϳ450 kcal). The study medica- history of type 2 diabetes, medical his- 0.11% in the metformin plus miglitol

990 DIABETES CARE, VOLUME 24, NUMBER 6, JUNE 2001 Chiasson and Naditch

Table 1—Demographic and clinical data on randomized patients compared with metformin monotherapy (34.6%). Parameters Placebo MIG MET MIG ϩ MET Reductions in levels of postprandial plasma glucose were observed in all the n 83 82 83 76 active treatment groups, in contrast to an Age (years) 57.7 Ϯ 9.9 57.3 Ϯ 9.0 57.9 Ϯ 8.6 58.9 Ϯ 7.9 increase in patients taking placebo (Table Weight (kg) 88.6 Ϯ 14.1 91.0 Ϯ 15.5 89.0 Ϯ 17.8 85.6 Ϯ 13.1 2). The reductions in patients receiving Race metformin plus miglitol combination Caucasian 76 (91.6) 73 (89.0) 73 (88.0) 70 (92.1) therapy were significantly greater (P Ͻ Black 1 (1.2) 0 (0.0) 1 (1.2) 3 (3.9) 0.0001) than those in patients on met- Asian 4 (4.8) 4 (4.9) 6 (7.2) 3 (3.9) formin monotherapy. The reduction in Other 2 (2.4) 5 (6.1) 3 (3.6) 0 (0.0) fasting plasma glucose was also greater in 2 Ϯ Ϯ Ϯ Ϯ BMI (kg/m ) 31.1 4.4 31.1 4.5 30.7 5.1 29.5 3.8 patients receiving metformin plus migli- Male/female 56/27 64/18 61/22 59/17 tol combination therapy than in patients Ϯ Ϯ Ϯ Ϯ Duration of diabetes (years) 5.1 4.9 5.2 4.7 7.5 7.4 6.1 5.5 receiving metformin monotherapy (P ϭ Previous use of oral 0.0025). hypoglycemics agents Although the mean change in post- None 48 (57.8) 42 (51.2) 55 (66.3) 46 (60.5) prandial plasma insulin at 60 min did not Metformin 22 (26.5) 18 (22.0) 19 (22.9) 16 (21.1) reach statistical significance for the com- Sulphonylureas 33 (39.8) 25 (30.5) 43 (51.8) 35 (46.1) parison between metformin plus miglitol HbA (%) 8.1 Ϯ 0.7 8.2 Ϯ 0.9 8.2 Ϯ 0.9 8.3 Ϯ 0.8 lc combination therapy and metformin Data are means Ϯ SD and n (%). MIG, miglitol; MET, metformin. monotherapy, a statistically significant difference was observed in favor of the combination group (P ϭ 0.002, compar- combination. The latter group achieved combination therapy at 90 and 120 min Ͻ ϭ ing miglitol plus metformin and met- the targeted HbA1c level of 7.0% recom- (P 0.0143 and 0.0177, respectively) formin monotherapy). The placebo- mended by the American Diabetes Asso- (Table 2). Postprandial plasma insulin substracted mean change in HbA1c or the ciation (20). Furthermore, significantly levels decreased more in the placebo actual treatment effect is illustrated in Fig. (P ϭ 0.0014) more patients were classi- group than in the metformin group, Ն 1. The mean reduction in HbA1c com- fied as responders (i.e., showed 15% re- which could be caused by bias in the pa- pared with placebo was –0.37% for migli- duction from baseline in HbA1c or tients’ disease status. Because the patients Ϫ Ͻ tol treatment, 1.25% for metformin achieved HbA1c 7.0%) to metformin were not newly diagnosed, their duration treatment, and –1.78% for metformin plus miglitol combination therapy of disease varied considerably, and some plus miglitol treatment. The end-of- (70.6%) compared with metformin had previously been taking sulfonylurea, Ϯ treatment mean SEM of HbA1c was monotherapy (45.5%); even if only which stimulates release of pancreatic in- Ϯ Ϯ Ͻ 8.5 0.1% for placebo, 8.2 0.2% for HbA1c 7.0% was used to define re- sulin. It is clear that lower postprandial miglitol, 7.3 Ϯ 0.1% for metformin, and sponders, combination therapy still plasma insulin levels despite higher post- 6.9 Ϯ 0.1% for metformin plus miglitol maintained a high responder rate (64.0%) prandial plasma glucose levels suggests

Table 2—Mean change from baseline in selected study variables (ITT population)

MET versus Parameters Placebo MIG MET MIG ϩ MET MIG ϩ MET (P) n 82 80 81 75 Ϯ Ϯ Ϫ Ϯ Ϫ Ϯ HbAlc (%) 0.38 0.12 0.02 0.10 0.85 0.12 1.39 0.11 0.002 Fasting plasma glucose (mg/dl) Ϫ0.9 Ϯ 5.7 Ϫ1.0 Ϯ 5.6 Ϫ20.4 Ϯ 5.1 Ϫ44.8 Ϯ 5.1 0.0025 Postprandial plasma glucose (mg/dl) 60 min 6.2 Ϯ 7.8 Ϫ16.3 Ϯ 7.6 Ϫ28.7 Ϯ 7.0 Ϫ81.9 Ϯ 8.1 Ͻ0.0001 90 min 6.9 Ϯ 8.3 Ϫ24.1 Ϯ 8.5 Ϫ33.7 Ϯ 7.7 Ϫ95.4 Ϯ 9.1 Ͻ0.0001 120 min 4.0 Ϯ 8.5 Ϫ23.2 Ϯ 8.5 Ϫ37.9 Ϯ 7.9 Ϫ86.4 Ϯ 9.3 Ͻ0.0001 Incremental plasma glucose AUC (mg ⅐ h/dl) 7.9 Ϯ 7.0 Ϫ34.7 Ϯ 7.2 Ϫ18.0 Ϯ 7.3 Ϫ59.0 Ϯ 7.2 Fasting plasma insulin (pmol/l) Ϫ0.3 Ϯ 14.0 Ϫ18.5 Ϯ 14.0 Ϫ17.4 Ϯ 4.7 Ϫ12.4 Ϯ 3.8 NS Postprandial plasma insulin (pmol/l) 60 min Ϫ48.8 Ϯ 18.0 Ϫ68.2 Ϯ 18.1 Ϫ19.9 Ϯ 15.0 Ϫ33.5 Ϯ 14.4 NS 90 min Ϫ40.3 Ϯ 17.4 Ϫ88.4 Ϯ 18.6 Ϫ18.9 Ϯ 14.8 Ϫ81.6 Ϯ 18.1 0.0143 120 min Ϫ48.4 Ϯ 19.0 Ϫ63.6 Ϯ 19.8 3.9 Ϯ 14.8 Ϫ59.1 Ϯ 19.3 0.0177 Incremental plasma insulin AUC (pmol ⅐ h/l) Ϫ45.0 Ϯ 23.1 Ϫ72.8 Ϯ 21.9 0.5 Ϯ 17.7 Ϫ60.1Ϯ 19.4 0.0592 Body weight (kg) Ϫ0.69 Ϯ 0.27 Ϫ0.42 Ϯ 0.29 Ϫ0.79 Ϯ 0.33 Ϫ1.87 Ϯ 0.33 NS Data are mean Ϯ SE. NS, not significant, MIG, miglitol, MET, metformin.

DIABETES CARE, VOLUME 24, NUMBER 6, JUNE 2001 991 Miglitol plus metformin in type 2 diabetes

and 19 (25.0%) in the metformin plus miglitol combination therapy group. Flat- ulence and diarrhea were the most com- mon adverse events associated with premature discontinuation from the study. A total of 16 patients reported 18 se- rious adverse events postrandomization: 3 of these events occurred in the placebo group, 2 in the miglitol group, 4 in the metformin group, and 9 in the metformin plus miglitol combination group. None of these serious adverse events were deemed by the investigators to be either probably or possibly related to the study drug. No deaths occurred during this study. No se- vere hypoglycemic episodes were re- ported. The rate of hypoglycemia was slightly higher in patients receiving met- formin plus miglitol combination therapy (13.2 vs. 9.6% receiving metformin), but this difference was not clinically significant. Figure 1—The mean absolute change (open bar) and the placebo-subtracted change (closed bar) Treatment groups did not show any in HbA1c from baseline to end of treatment for miglitol monotherapy, metformin monotherapy, and clinically or statistically significant differ- miglitol plus metformin combination therapy. ences in hematological and biochemical parameters or urinalysis. One patient in ␤ the metformin group showed an elevation decreased -cell function. This could be with 74 of 76 (97.4%) on metformin plus Ͼ1.8 times the upper limit of normal in caused by glucotoxicity resulting from in- miglitol combination therapy (Table 3), alanine aminotransferase, but this was not creased plasma glucose in the placebo which is an increased incidence of 3.4%. considered clinically relevant. Glycosuria group (21). Changes in fasting serum in- For comparison, the incidence was 85.5% was observed considerably less frequent- sulin and triglyceride levels observed in the placebo group and 96.3% in the ly for the metformin plus miglitol com- from baseline to the end of treatment did miglitol group. Most adverse events were bination group, with an incidence of 7% not differ significantly between met- gastrointestinal; flatulence and diarrhea compared with 26% for metformin formin plus miglitol combination therapy were reported more frequently by pa- monotherapy. There was no significant and metformin monotherapy and showed tients receiving miglitol or metformin laboratory abnormality or change in vital no consistent trend. plus miglitol combination therapy than signs during the study. Although all treat- A total of 324 patients were valid for those receiving the other treatments. ment groups showed a mean decrease in the safety analysis. After randomization, A total of 37 (11.4%) randomized pa- body weight, intergroup comparison sug- 302 (93.2%) patients reported adverse tients discontinued the study prematurely gested that patients who received met- events. The proportion of patients expe- because of adverse events: 2 (2.4%) in the formin plus miglitol combination therapy riencing adverse events in the metformin placebo group, 11 (13.4%) in the miglitol lost more weight (Table 2) than those in group was 78 of 83 (94.0%), compared group, 5 (6.0%) in the metformin group, the other treatment groups: Ϫ1.87 kg for metformin plus miglitol combination Table 3—Incidence of most common adverse events compared with –0.79 kg for metformin alone, –0.42 kg for miglitol alone, and Ϫ0.69 kg for placebo (P ϭ NS). Placebo MIG MET MIG ϩ MET n 83 82 83 76 Any adverse events 71 (85.5) 79 (96.3) 78 (94.0) 74 (97.4) CONCLUSIONS — Previous studies Digestive system have shown the efficacy and safety of Any event 29 (34.9) 58 (70.7) 50 (60.2) 66 (86.8) miglitol as monotherapy and in combina- Flatulence 12 (14.5) 46 (56.1) 24 (28.9) 48 (63.2) tion with sulfonylureas in type 2 diabetes Diarrhea 9 (10.8) 35 (42.7) 23 (27.7) 42 (53.3) (11,13,18,20,22,23). The present study Constipation 5 (6.0) 5 (6.1) 7 (8.4) 7 (9.2) demonstrates that miglitol in combina- Nausea 2 (2.4) 7 (8.5) 14 (16.9) 13 (17.1) tion with metformin provides a synergis- Dyspepsia 2 (2.4) 6 (7.3) 7 (8.4) 11 (14.5) tic effect on glycemic control, as indicated Abdominal cramps 2 (2.4) 4 (4.9) 5 (6.0) 6 (7.9) by the marked reductions in HbA1c and Hypoglycemia 7 (8.4) 7 (8.5) 8 (9.6) 10 (13.2) plasma glucose levels in middle-aged Data are n (%). Common adverse events are considered those with incidence Ͼ7% in at least one treatment patients in whom type 2 diabetes is insuf- group. MIG, miglitol, MET, metformin. ficiently controlled by dietary manage-

992 DIABETES CARE, VOLUME 24, NUMBER 6, JUNE 2001 Chiasson and Naditch ment. Combination therapy produced glucose production and release and indi- ication. It is possible that slower titration significantly greater reductions in HbA1c, rectly by increasing peripheral tissue sen- of miglitol would reduce the gastrointes- fasting plasma glucose, and particularly sitivity to insulin (27); miglitol acts at the tinal side effects and improve compliance, postprandial plasma glucose than met- small intestine by delaying the digestion resulting in better improvement in HbA1c. formin monotherapy. The mean HbA1c at of complex carbohydrates (28). Interest- Miglitol in combination with met- the end of treatment of patients receiving ingly, patients receiving metformin plus formin was found to be safe and well tol- metformin was 7.3%, which is slightly miglitol combination therapy tended to erated in the study cohort. The incidence Ͻ higher than the goal HbA1c level of 7% lose more weight than those on the other of gastrointestinal adverse events was given by current clinical practice recom- treatment regimens (4). Weight loss may higher in the treatment groups receiving mendations (20), whereas patients on indirectly improve glycemic control by miglitol; this is in line with the known metformin plus miglitol combination decreasing insulin resistance (29), which mechanism of action of an ␣-glucosidase therapy achieved the treatment goal with might have been another positive con- inhibitor. It is possible that slower titra- a mean end-of-treatment HbA1c level of tributory factor to the superiority of the tion of the drug, as with acarbose, could 6.9%. The greater reduction in HbA1c in metformin plus miglitol combination diminish the incidence of gastrointestinal patients on the metformin plus miglitol therapy. Although combination therapy side effects (32). Overall, the incidence of combination therapy is therefore clini- did not have any significant effect on fast- side effects with metformin plus miglitol cally significant in achieving the treat- ing plasma insulin compared with met- combination therapy was not signifi- ment goal. The clinical importance of this formin alone (Ϫ12.4 Ϯ 3.8 vs. Ϫ17.4 Ϯ cantly different from that observed for ei- is supported by the U.K. Prospective Di- 4.7 pmol/l; P ϭ NS), it did have a ten- ther miglitol or metformin monotherapy, abetes Study data, which indicates that for dency to decrease the postprandial incre- although the rate of discontinuations was every 1% reduction in HbA1c, there is a mental plasma insulin area under the higher with combination therapy. Only a 35% reduction in risk of microvascular curve (AUC) (Ϫ60.1 Ϯ 19.4 vs. ϩ0.5 Ϯ trend toward an increase in the number of complications (24). The superiority of the 17.7 pmol ⅐ h/l; P ϭ 0.059). This is prob- gastrointestinal side effects in the met- combination treatment was also demon- ably caused by the effect of miglitol be- formin plus miglitol combination therapy strated by the higher response rate of pa- cause miglitol alone resulted in a decrease group was observed. There was no evi- tients on the metformin plus miglitol in postprandial plasma insulin AUC of dence of any serious adverse events asso- combination therapy than on mono- Ϫ72.8 Ϯ 21.9 pmol ⅐ h/l. The reduction ciated with this combination regimen. No therapy. These observations support in postprandial plasma glucose and insu- severe hypoglycemic episodes were ob- those of Mooradian in elderly type 2 dia- lin could result in improved insulin sen- served in this study, and the incidence of betic subjects (18). sitivity, as we have shown in elderly type mild-to-moderate hypoglycemia was low Postprandial hyperglycemia is recog- 2 diabetic subjects with acarbose treat- and comparable across all study treat- nized as an independent risk factor for ment (30). ments. The lack of deleterious effects on macrovascular complications (25,26). Monotherapy with miglitol alone re- liver enzymes indicates that regular mon- However, the normalization of postpran- sulted in a treatment effect of Ϫ0.37% in itoring of liver function during combina- dial plasma glucose peaks in clinical prac- HbA1c compared with placebo (Fig. 1). tion therapy with miglitol and metformin tice is recognized as more problematic This is a much smaller effect than that is not required. than the overall management of fasting observed in previously reported studies From the present study, it can be con- plasma glucose levels. Because ␣-glucosi- (11,13,22,23,31), in which a reduction in cluded that miglitol, the first pseudo- dase inhibitors have been reported previ- HbA1c between 0.74 and 1.19% was monosaccharide ␣-glucosidase inhibitor, ously to smooth and lower postprandial shown. It is interesting that Johnston et al. can be combined effectively with met- plasma glucose peaks (8–10), miglitol in (11) could not show any better efficacy of formin therapy to give significantly combination with metformin does offer miglitol 100 mg three times a day com- greater reductions in HbA1c and post- an advantage in achieving the important pared with 50 mg three times a day. How- prandial plasma glucose levels than met- goal of postprandial glucose management ever, postprandial plasma glucose was formin alone, with a good safety profile, as well as overall glycemic control, espe- lower after 100 mg three times a day on in patients in whom type 2 diabetes is cially for patients with postprandial hyper- the test meal. It is possible that the discor- insufficiently controlled by diet alone. glycemia refractory to other treatments. dance between HbA1c and postprandial Miglitol and metformin may act synergis- The reductions in fasting and post- plasma glucose was caused by the lower tically to confer this additional glycemic prandial plasma glucose as well as in compliance at 100 mg three times a day control, especially on postprandial HbA1c in response to metformin plus because of increased gastrointestinal side plasma glucose peaks, and may thereby miglitol combination therapy were effects. Similarly, in the present study, the help to reduce the risk of microvascular greater than the added effects of both effect of miglitol on incremental post- and macrovascular diabetic complications. miglitol and metformin alone. It is sug- prandial plasma glucose AUC during the gested that the administration of the two test meal was significantly greater than drugs together has a synergistic effect on with metformin alone (Ϫ34.7 Ϯ 7.2 vs. Ϫ Ϯ ⅐ Ͻ Acknowledgments— This study was funded glycemic control; this is plausible because 18.0 7.3 mg h/dl; P 0.001), con- by an unrestricted research grant from Bayer miglitol and metformin reduce plasma firming the efficacy of the drug. Again, the Canada, Inc., and additional financial support glucose levels through completely differ- discordance between HbA1c and post- was received from Sanofi-Synthelabo. ent mechanisms of action. Metformin acts prandial plasma glucose could be caused We thank the coordinating nurses and die- on the liver directly by decreasing hepatic by poor compliance with the study med- ticians in all centers for their contribution.

DIABETES CARE, VOLUME 24, NUMBER 6, JUNE 2001 993 Miglitol plus metformin in type 2 diabetes

This study was presented as a poster at the emergence of ␣-glucosidase inhibitors. 1):S32–S41, 1999 60th annual scientific sessions of the American Drugs 44 (Suppl. 3):21–28, 1992 21. Rossetti L, Giaccari A, DeFronzo RA: Glu- Diabetes Association, San Antonio, Texas, 9. Holman RR, Steemson J, Turner RC: Post- cose toxicity. Diabetes Care 13:610–630, 9–13 June 2000. prandial glycaemic reduction by an ␣-glu- 1990 cosidase inhibitor in type 2 diabetic patients 22. Johnston PS, Feig PU, Coniff RF, Krol A, APPENDIX with therapeutically attained basal normo- Kelley DE, Mooradian AD: Chronic treat- glycaemia. Diabetes Res 18:149–153, 1991 ment of African-American type 2 diabetic The Miglitol University Canadian Investi- 10. Heinz G, Komjati M, Korn A, Waldhausl patients with ␣-glucosidase inhibition. gator Group included: Dr. Hertzel C. Ger- W: Reduction of postprandial blood glu- Diabetes Care 21:416–422, 1998 stein, McMaster University Medical cose by the alpha-glucosidase inhibitor 23. Johnston PS, Feig PU, Coniff RF, Krol A, Center; Dr. Robert J. Josse, St. Michael’s Miglitol (BAY m 1099) in type II diabetes. Davidson JA, Haffner SM: Long-term ti- Hospital; Dr. David Lau, Ottawa Civic Eur J Clin Pharmacol 37:33–36, 1989 trated-dose ␣-glucosidase inhibition in Hospital; Dr. Lawrence A. Leiter, St. Mi- 11. Johnston PS, Coniff RF, Hoogwerf BJ, non-insulin-requiring Hispanic NIDDM chael’s Hospital; Dr. Ruth McManus and Santiago JV, Pi-Sunyer FX, Krol A: Effects patients. Diabetes Care 21:409–415, 1998 Dr. N. Wilson Rodger, St. Joseph’s Health of the carbohydrase inhibitor miglitol in 24. Stratton IM, Adler AI, Neil HA, Matthews Center; Dr. Thomas M.S. Wolever, Uni- sulfonylurea-treated NIDDM patients. Di- DR, Manley SE, Cull CA, Hadden D, abetes Care 17:20–29, 1994 versity of Toronto, Toronto, ON; Dr. John Turner RC, Holman RR: Association of 12. Spengler M, Hansel G, Boehme K: Ef- glycaemia with macrovascular and micro- A. Hunt, Lion’s Gate Hospital; Dr. Hugh ficacy of 6 months monotherapy with D. Tildesley, Vancouver, BC; Dr. Pierre vascular complications of type 2 diabetes glucosidase inhibitor Acarbose versus (UKPDS 35): prospective observational Maheux, Center Universitaire de Sante´de sulphonylurea glibenclamide on meta- study. BMJ 321:405–412, 2000 l’Estrie; Dr. Jean-Franc¸ois Yale, McGill bolic control of dietary treated type II 25. Hanefeld M, Fischer S, Julius U, Schulze J, Nutrition Center, Quebec; PQ; Dr. Liam diabetics (NIDDM). Horm Metab Res Schwanebeck U, Schmechel H, Ziegel- Murphy, University of Manitoba, Win- (Suppl.) 26:50–51, 1992 asch HJ, Lindner J, and the DIS Group: nipeg, MB; Dr. Stuart A. Ross, Calgary 13. Segal P, Feig PU, Schernthaner G, Ratz- Risk factors for myocardial infarction and Metabolic Education and Research Cen- mann KP, Rybka J, Petzinna D, Berlin C: death in newly detected NIDDM: the Di- ter; Dr. Edmond A. Ryan, Heritage Medi- The efficacy and safety of miglitol therapy abetes Intervention Study, 11-year fol- compared with glibenclamide in patients cal Research Center, Edmonton, AB; and low-up. Diabetologia 39:1577–1583, 1996 with NIDDM inadequately controlled by 26. Coutinho M, Gerstein HC, Wang Y, Yusuf Dr. Makram A. Boctor, Royal University diet alone. Diabetes Care 20:687–691, S: The relationship between glucose and Hospital, Saskatoon, SK. 1997 incident cardiovascular events: a metar- 14. Escobar-Jimenez F, Barajas C, De Leiva A, egression analysis of published data from Cano FJ, Masoliver R, Herrera-Pombo JL, 20 studies of 95,783 individuals followed References Hernandez-Mijares A, Pinon F, de la Calle for 12.4 years. Diabetes Care 22:233–240, 1. U.K. Prospective Diabetes Study Group: H, Tebar J, Soler J, Cobos A, Guardiola E, 1999 Intensive blood-glucose control with sul- and the Miglitol Collaborative Group: Ef- phonylureas or insulin compared with ficacy and tolerability of miglitol in the 27. Inzucchi SE, Maggs DG, Spollett GR, Page conventional treatment and risk of com- treatment of patients with non-insulin- SL, Rife FS, Walton V, Shulman GI: Effi- plications in patients with type 2 diabetes dependent diabetes mellitus. Curr Ther cacy and metabolic effects of metformin (UKPDS 33). Lancet 352:837–853, 1998 Res 56:258–268, 1995 and troglitazone in type II diabetes melli- 2. U.K. Prospective Diabetes Study Group: 15. Saunier B, Kilker RD Jr, Tkacz JS, Quaroni tus. N Engl J Med 338:867–872, 1998 Effect of intensive blood-glucose control A, Herscovics A: Inhibition of N-linked 28. Taylor RH, Barker HM, Bowey EA, Can- with metformin on complications in over- complex oligosaccharide formation by field JE: Regulation of the absorption of weight patients with type 2 diabetes (UK- 1-deoxynojirimycin, an inhibitor of pro- dietary carbohydrate in man by two new PDS 34). Lancet 352:854–865, 1998 cessing glucosidases. J Biol Chem 257: glucosidase inhibitors. Gut 27:1471– 3. DeFronzo RA, Goodman AM, and the 14155–14161, 1982 1478, 1986 Multicenter Metformin Study Group: 16. Scott LJ, Spencer CM: Miglitol: a review of 29. Goodpaster BH, Kelley DE, Wing RR, Efficacy of metformin in patients with its therapeutic potential in type 2 diabetes Meier A, Thaete FL: Effects of weight loss non-insulin-dependent diabetes mellitus. mellitus. Drugs 59:521–549, 2000 on regional fat distribution and insulin N Engl J Med 333:541–549, 1995 17. Pagano G, Marena S, Corgiat-Mansin L, sensitivity in obesity. Diabetes 48:839– 4. Johansen K: Efficacy of metformin in the Cravero F, Gioada C, Bozza M, Rossi CM: 847, 1999 treatment of NIDDM: meta-analysis. Dia- Comparison of miglitol and gliben- 30. Meneilly GS, Ryan EA, Radziuk J, Lau betes Care 22:33–37, 1999 clamide in diet-treated type 2 diabetic pa- DCW, Yale J-F, Morais J, Chiasson J-L, 5. Campbell LK, White JR, Campbell RK: tients. Diabete Metab 21:162–167, 1995 Rabasa-Lhoret R, Maheux P, Tessier D, Acarbose: its role in the treatment of dia- 18. Mooradian AD: Drug therapy of non-in- Wolever T, Josse RG, Elahi D: Effect of betes mellitus. Ann Pharmacother 30: sulin-dependent diabetes mellitus in the acarbose on insulin sensitivity in elderly 1255–1262, 1996 elderly. Drugs 51:931–941, 1996 patients with diabetes. Diabetes Care 23: 6. Bischoff H: Pharmacology of ␣-glucosi- 19. Franz MJ, Horton ES, Bantle JP, Beebe 1162–1167, 2000 dase inhibition. Eur J Clin Invest 24 CA, Brunzell JD, Coulston AM, Henry RR, 31. Lebovitz HE: ␣-glucosidase inhibitors as (Suppl. 3):3–10, 1994 Hoogwerf B, Stacpoole PW: Nutritional agents in the treatment of diabetes. Diabe- 7. Joubert PH, Venter HL, Foukaridis GN: principles for the management of diabetes tes Rev 6:132–145, 1998 The effect of miglitol and acarbose after an and related complications. Diabetes Care 32. May C: Efficacy and tolerability of step- oral glucose load: a novel hypoglycemic 17:490–518, 1994 wise increasing dosage of acarbose in mechanism? Br J Clin Pharmacol 30:391– 20. American Diabetes Association: Stan- patients with non-insulin-dependent dia- 396, 1990 dards of medical care for patients with di- betes (NIDDM), treated with sulphonylu- 8. Lebovitz HE: Oral antidiabetic agents: the abetes mellitus. Diabetes Care 22 (Suppl. reas. Diabetes Und Stoffwechsel 4:3–8, 1995

994 DIABETES CARE, VOLUME 24, NUMBER 6, JUNE 2001