Efficacy and Safety of Combination Therapy: Repaglinide plus metformin versus nateglinide plus metformin

Patients eligible for enrollment were adults (≥18 years old) who had type 2 diabetes for at least 3 months and BMI values of 24–42 kg/m². Subjects were stratified by baseline HbA_1c value (<9% or ≥9%). Enrolled patients had HbA_1c values >7% and ≤12% in previous monotherapy with a sulfonylurea (at ≥25% of the maximum dose), metformin (>1,000 mg/day), or low-dose Glucovance (glyburide ≤2.5 mg and metformin ≤500 mg).

This clinical trial was conducted in accordance with the provisions of the Declaration of Helsinki for participation of subjects in human research. The protocol received approval of relevant institutional review boards before initiation of any trial-related activities.

This clinical trial was a multicenter, randomized, parallel-group, open-label comparison of repaglinide plus metformin versus nateglinide plus metformin treatment for a period of 16 weeks.

Subjects previously treated with a sulfonylurea had a 4-week run-in period of metformin treatment (500 mg b.i.d. for 2 weeks, followed by metformin 1,000 mg b.i.d. for 2 weeks and thereafter doses taken with meals). Those previously treated with metformin or low-dose Glucovance received 1,000 mg metformin b.i.d. for 4 weeks. Subjects were then randomly assigned to addition of either mealtime repaglinide (n = 96) or mealtime nateglinide (n = 96) to their metformin regimen. Repaglinide or nateglinide was to be administered 1–30 min before mealtimes. Dose titration was carried out according to product labeling of each secretagogue. During a 2-week titration period, the dosage of repaglinide was increased stepwise from 1.0 to 2.0 and to 4.0 mg per meal at weekly visits based on the results of 8-point self-monitoring of blood glucose (SMBG) (maximum dose, 16 mg/day). Targets for glycemic control during the 2-week dose titration period were SMBG preprandial values of 80–140 mg/dl. Starting nateglinide dose was 120 mg per meal (the maximum daily dose), which could be reduced to 60 mg/meal in response to hypoglycemia episodes. An additional 14 weeks of maintenance therapy followed, during which secretagogue dosage adjustment was still possible as needed.

Liquid meal challenge testing (two cans of Boost = 480 kcal) was performed at baseline and the 16-week visit. Plasma glucose (hexokinase assay), insulin (immunometric assay), and glucagon (radioimmunoassay) levels were determined by sampling for up to 4 h postchallenge (assays by Icon Laboratories). At the 16-week visit, repaglinide or nateglinide was given 10 min before the liquid test meal.

The primary efficacy end points of this trial were final HbA_1c values (HPLC assay, Icon Laboratories; HbA_1c values for nondiabetic individuals, 4.7–6.4%) and changes in HbA_1c from baseline. Secondary efficacy end points included fasting plasma glucose (FPG), and assessment of glucose area under the time-concentration curves from 0 to 240 min (AUC_{0–240 min}), insulin AUC_{0–240 min}, and glucagon AUC_{0–240 min} after a liquid test meal (at baseline and at end of study).

Adverse events and reports of hypoglycemic episodes were recorded at all study visits. Subjects were asked to conduct SMBG from the randomization visit onward. For the purposes of this clinical trial, hypoglycemic episodes were defined as follows. Major hypoglycemic episodes were events having severe central nervous system symptoms consistent with hypoglycemia in which the subject was unable to treat him- or herself, having blood glucose readings <50 mg/dl and/or reversal of symptoms by treatment (food intake, glucagon, or intravenous glucose). Minor hypoglycemic episodes included events with symptoms that were consistent with hypoglycemia symptoms and confirmed blood glucose levels <50 mg/dl, or a blood glucose level <50 mg/dl, even if there were no symptoms of hypoglycemia.

In the event of patient withdrawal or missing data after baseline, missing values of HbA_1c and FPG were substituted by imputed data (calculated by the incremental mean imputation [IMI] method) (9). Appropriate simulations have indicated that the IMI method is more precise than the last observation carried forward (LOCF) method for datasets resembling this clinical trial (9). The IMI results were used because this method is considered to be more precise and more conservative for comparison of treatment groups. Differences between the combination therapy groups in the change in HbA_1c or FPG values were compared by ANOVA (with and without adjustment for baseline imbalance).

It was projected that 64 randomized patients per group who had at least one postdosing efficacy evaluation would be required to detect 0.6% units of difference in HbA_1c values with 80% power. Liquid meal testing postprandial values of glucose, insulin, and glucagon were analyzed in terms of AUC_{0–240 min}, where these values were adjusted for the baseline level at time 0 (meal initiation).

Demographic characteristics and baseline values of enrolled patients are summarized in Table 1. The treatment groups were generally comparable in such variables as duration of diabetes, previous diabetes treatment, age, and BMI. The repaglinide group had a somewhat higher number of women and persons of Hispanic ethnic background than the nateglinide treatment group. Inclusion of race or sex as variables in this model indicated that these variables were associated with no significant differences in response to treatment.

The primary efficacy parameter of this study, HbA_1c, showed significant divergence between the treatment groups from week 4 onward (Fig. 1A). Final HbA_1c values were lower for repaglinide/metformin combination therapy than nateglinide/metformin treatment (Table 2). Mean end-of-study changes in HbA_1c values from baseline were significantly greater for the repaglinide/metformin combination regimen than for nateglinide/metformin (−1.28 vs. −0.67%; P < 0.001). The percentage of patients who achieved final HbA_1c values ≤7% was also higher for repaglinide/metformin therapy (59 vs. 47% for nateglinide/metformin). Of the 47 repaglinide-treated patients having initial HbA_1c values >8%, there were 17 (36%) who achieved final HbA_1c values ≤7%. By contrast, only 7 of 34 (21%) comparable nateglinide-treated patients showed the same improvement in final HbA_1c values.

FPG values were significantly different for the two treatment groups within a week of therapy, with this treatment difference persisting to the end of the study (Fig. 1B). FPG values reached a steady state after ∼4 weeks of therapy. Mean end-of-study reductions of FPG values from baseline were significantly greater for the repaglinide/metformin group (−39 vs. −21 mg/dl for nateglinide/metformin; P = 0.002) (Table 2). At the end of the 16-week maintenance therapy, 48% of repaglinide/metformin group patients had reductions of FPG values by >40 mg/dl, whereas only 26% of nateglinide/metformin group patients had a glycemic response of this magnitude.

The mean 8-point SMBG profiles of the two treatment groups are summarized in Fig. 2. Repaglinide/metformin therapy was associated with mean SMBG values that were consistently lower than those of the nateglinide/metformin group at all times measured. Such treatment-related differences were significant at several times during the day: before breakfast, before lunch, and at 2:00 a.m.

At the end of 16-week treatment period, liquid meal challenge testing demonstrated slightly lower postmealtime glucose levels in the repaglinide/metformin group than in the nateglinide/metformin group (Fig. 3A). Mean end-of-study reductions in postprandial glucose levels from baseline were not significantly different between the groups (glucose AUC_{0–240 min} in Table 2). The treatments were also comparable for changes in insulin AUC_{0–240 min} and glucagon AUC_{0–240 min} during the study (Fig. 3B and Table 2).

Most nateglinide-treated patients did not require dose reduction from the maximal recommended dose to prevent hypoglycemic events, whereas most repaglinide-treated patients did not reach maximal recommended doses. At the end of 16 weeks of treatment, median final doses of mealtime secretagogues were 5.0 mg/day for repaglinide and 360 mg/day for nateglinide. For the nateglinide/metformin group, 82% of patients received the daily maximal dose of nateglinide, whereas only 7% of repaglinide-treated patients received the maximal daily dosage. In both groups, the median dose of metformin was 2,000 mg/day.

The fraction of patients who discontinued study treatment was slightly higher for the nateglinide/metformin regimen than for the repaglinide/metformin regimen (18 vs. 7%, respectively) (Table 1). The single most frequently cited reason for the higher rate of discontinuation of therapy among nateglinide/metformin group patients was lack of efficacy, as assessed by the investigator (Table 1).

There were no patients in either treatment group who experienced major hypoglycemic episodes (requiring the assistance of another person). Minor hypoglycemic episodes occurred in 7% of the patients of the repaglinide/metformin group compared with 2% of the patients in the nateglinide/metformin group. Most of these minor hypoglycemic events were accompanied by symptoms (five of seven events for repaglinide/metformin and all events for nateglinide/metformin). The most frequent adverse event in both groups was upper respiratory tract infection (21% of repaglinide/metformin group vs. 12% of nateglinide/metformin group). Adverse events occurring in 3–8% of patients in both groups included nausea, viral infection, accidental injury, sinusitis, diarrhea, and headache. The repaglinide/metformin group had 5% incidence of arthralgia and 5% incidence of chest pain, as compared with 1% for each in the nateglinide/metformin group. In general, the treatment groups showed no noteworthy differences in safety measures during 16 weeks of therapy.

Both repaglinide and nateglinide showed small weight changes from baseline during 16 weeks of therapy (mean changes were 0.6-kg gain vs. 0.5-kg loss, respectively).

Repaglinide and nateglinide are insulin secretagogues that are administered at mealtimes, and both are approved in the U.S. for use in a combination regimen with metformin. No controlled clinical trial has been previously reported that would provide a direct comparison of efficacy and safety of repaglinide/metformin versus nateglinide/metformin.

The repaglinide/metformin combination therapy regimen rapidly reduced FPG values to a steady state in 4 weeks and had a stable effect on HbA_1c values by 12 weeks. Reductions of HbA_1c values differed by ∼0.6% between treatment groups, a difference that is clinically and statistically significant (P < 0.001). From the results of this clinical trial, it would be estimated that ∼22% more patients achieved FPG reductions of >40 mg/dl by using repaglinide/metformin rather than nateglinide/metformin. A lesser efficacy response for nateglinide/metformin therapy cannot be attributed to dosage of the mealtime secretagogues: median final daily dosage of repaglinide was 5.0 mg/day (only 31% of recommended maximum dose), whereas the median daily dosage of nateglinide was 360 mg/day (100% of the recommended maximum dose). Most patients did not reach the maximal recommended dose of repaglinide during dose titration, for the most part because glycemic goals were achieved. The dose adjustment regimen (based on product labeling) dictated that all nateglinide doses began at the maximal dose, while repaglinide was titrated upward from 1 mg/meal. It is possible that differences in efficacy of repaglinide and nateglinide would be even greater under conditions where both agents were used at the maximal dosage.

The efficacy of repaglinide/metformin observed in the current study is consistent with earlier clinical trials of this combination therapy (5). For nateglinide/metformin therapy, HbA_1c values were reduced by 0.67% relative to baseline in 16 weeks, whereas some studies have reported reductions of as much as 1.4% in 24 weeks (3). However, such differences may reflect study design, since the latter data were collected after a 2-month oral antidiabetic (OAD) washout period that would have resulted in a significant rise in HbA_1c values before initiation of combination therapy.

In assays following a single liquid test meal, the two treatments produced similar reductions of postprandial glucose peaks, stimulation of insulin, and effects on glucagon levels. This clinical trial did not detect any evidence that nateglinide has clinically significant differences in effects on the earliest stages of insulin secretion. The lesser clinical efficacy of nateglinide in the current trial (as measured by HbA_1c and FPG levels) may be a result of its lower affinity for the β-cell molecular target (half-maximal inhibitory concentration [IC₅₀] for K_ATP channels = 7.4 μmol/l for nateglinide vs. 5 nm for repaglinide) (10). It has been reported that the duration of repaglinide inhibitory actions at K_ATP channels is notably longer than that of nateglinide (10), indicating there may be small differences in the late postprandial actions of these two agents. Although such differences may not be apparent in a single meal challenge test, they may have cumulative impact during repeated dosing.

Safety parameters did not show any notable differences between the two treatments. There were no notable differences in weight changes, laboratory values, or hypoglycemic event frequency for repaglinide/metformin and nateglinide/metformin.

The current study provides a direct comparison that will obviate the need to draw comparisons between separate and potentially disparate clinical studies of the efficacy of the individual drugs. Such comparisons have previously implied that nateglinide has an efficacy that is somewhat less than sulfonylureas and repaglinide (11).

In conclusion, combination therapy of repaglinide plus metformin was a safe and effective therapy in the treatment of type 2 diabetes after unsatisfactory response to OAD monotherapy (sulfonylureas, metformin, low-dose Glucovance). The comparison regimen of nateglinide/metformin showed significantly less reduction of glycemic parameters, with comparable safety.

Mary Ann Banerji, Steven S. Bimson, Shari A. Brazinky, Dennis Buth, Leif Christiansen, James Clower III, James Cuellar, James Farrell, Susan Greco, David R. Hassman, Priscilla Hollander, C. Scott Horn, Edward Kerwin, Allen King, Leslie Klaff, Nelson Kopyt, Steven B. Leichter, Janet McGill, Joseph Milburn, Marcia Miller, Margaret Nunez, Andres Patron, Jorge Pino, Larry Popeil, Philip Raskin, Julio Rosenstock, Mohammed F. Saad, Don Schumacher, Stan Seagraves, Barry Seidman, Richard Sievers, Stephen A. South, Scott Touger, Richard Weiss, and Evan J. Zimmer.

This clinical trial was supported by a research grant from Novo Nordisk Pharmaceuticals.

Technical assistance: Julie D. Bau, Diane Bellew, Terri Burch, Dara Clark, Paula Clark, Marianna Cooper, Andrea DaCosta, Suresh Daniels, Pam DeBold, Julie Dela Cruz, Melanie S. Diaz, Melissa Dottery, Tara Fraser, Mary Ann Gergits, Lisa Givens, Krista Hill, Sujata Jinagouda, Kay King, Ellen Kreifels, Tracy Krepel, Beth Lehman, Kurt Malphurs, Regina McFadden, Ruthanne Mezner, Tatyana Noryan, Colleen Person, Christine Pool, Dorothy Pyatt, Kim Quiroga, Pam Schearer, Donal Schumacher, Eric Serfer, Chrison Sitton, Tammy Skelton, Lindsay Stanhope, Tammy Stitch, Barbara A. Sumpter, Shari Touger, Paula Wells, Nancy E. Wilson, and Gary Wolf.