Effects of Exenatide (Exendin-4) on Glycemic Control Over 30 Weeks in Sulfonylurea-Treated Patients With Type 2 Diabetes

Subjects were 22–76 years of age and had type 2 diabetes treated with at least the maximally effective dose of a sulfonylurea as monotherapy (defined below) for at least 3 months before screening. General inclusion criteria were a screening fasting plasma glucose concentration <240 mg/dl, BMI 27–45 kg/m², and HbA_1c 7.1–11.0%, inclusive. In addition, subjects had stable weight (±10%) for 3 months before screening and had no clinically relevant (for a type 2 diabetic population) abnormal laboratory test values (>25% outside normal laboratory values). Female subjects were postmenopausal or surgically sterile or using contraceptives for at least 3 months before screening and continuing throughout the study. Subjects were excluded if they had used metformin, thiazolidinediones, meglitinides, α-glucosidase inhibitors, exogenous insulin therapy, or weight-loss drugs within the prior 3 months. Further exclusion criteria included therapy with corticosteroids, drugs known to affect gastrointestinal motility, transplantation medications, or any investigational drug. Subjects were excluded if they had evidence of clinically significant comorbid conditions.

Three hundred seventy-seven adults with sulfonylurea-treated type 2 diabetes participated at 101 sites in the U.S. (February 2002 to August 2003). Data from 100 sites were used in statistical analyses (1 site was closed during study conduct due to protocol noncompliance). A common clinical protocol was approved for each site by an institutional review board and in accordance with the principles described in the Declaration of Helsinki, including all amendments through the 1996 South Africa revision (13). All subjects provided written informed consent before participation.

This was a balanced, randomized, triple-blind, placebo-controlled, parallel-group, pivotal clinical study designed after consultation with the U.S. Food and Drug Administration to evaluate glycemic control, as assessed by HbA_1c, and safety. The study commenced with a 4-week, single-blind, lead-in period with subcutaneous injection of placebo twice daily. Thereafter, subjects were randomized to one of four treatment arms. Nausea had been the most frequent treatment-emergent adverse event in earlier clinical trials, but gradual dose escalation has been shown to attenuate this side effect (14). Therefore, the present study design included an acclimation period (4 weeks) at a lower exenatide fixed dose (5 μg b.i.d.) in treatment arms A and B, before the fixed dose of exenatide was either increased to 10 μg b.i.d. (arm B) or remained at 5 μg b.i.d. (arm A) for the duration of the study. Equivalent volumes of placebo to those administered to arms A and B were administered in treatment arms C and D. Study medication was self-injected subcutaneously in the abdomen within 15 min before meals in the morning and evening.

In an effort to standardize sulfonylurea use at study initiation, if required, subjects had their sulfonylurea dose adjusted before the placebo lead-in period to the maximally effective dose (4 mg/day glimepiride, 20 mg/day glipizide, 10 mg/day glipizide XL, 10 mg/day glyburide, 6 mg/day micronized glyburide, 350 mg/day chlorpropamide, or 500 mg/day tolazamide) (15–17). To address the risk of hypoglycemia, the protocol recommended progressive 50% reductions in sulfonylurea dose, eventual discontinuation (depending on the recurrence of hypoglycemia) in the event of a documented episode of hypoglycemia (glucose <60 mg/dl), or two undocumented but suspected episodes of hypoglycemia.

Any subject with either an HbA_1c change of 1.5% from baseline at any clinic visit before study termination or an HbA_1c ≥11.5% at week 18 or 24 could be withdrawn from the study (loss of glucose control). Similarly, subjects could be withdrawn if they had fasting plasma glucose values >240 mg/dl on two consecutive study visits or consistently recorded finger-stick fasting blood glucose values >260 mg/dl for at least 2 weeks, not secondary to a readily identified illness or pharmacological treatment.

Primary objectives were to evaluate glycemic control, primarily as assessed by HbA_1c, and safety. Secondary objectives included examining the effects of exenatide on fasting plasma glucose concentrations, body weight, and fasting concentrations of circulating insulin, proinsulin, and lipids. Safety end points included adverse events, clinical laboratory tests, physical examination, 12-lead electrocardiogram, and vital signs. Treatment-emergent adverse events were defined as those occurring upon or after receiving the first randomized dose. The emergence of anti-exenatide antibodies was also assessed.

Randomization was stratified according to screening HbA_1c values (<9.0% and ≥9.0%) to achieve a balanced distribution of subjects across treatment arms. A minimum sample size of 300 subjects who had at least one postbaseline HbA_1c measurement was estimated to provide ∼90% power to detect a difference of 0.6% in the change from baseline in HbA_1c values between at least one exenatide treatment arm and placebo (α = 0.05; Fisher’s protected testing procedure). Placebo arms C and D were combined for all analyses.

All inferential statistical tests were conducted at the significance level of 0.05 (two sided). A general linear model was used to test for differences in the change from baseline to each visit in HbA_1c across treatments (18,19). Factors in the model included treatment (placebo and two active treatment arms), strata of baseline HbA_1c (<9.0% and ≥9.0%), and study site as fixed effects. Before data analysis, sites were pooled according to geographic location to prevent the loss of too many degrees of freedom in the model. This pooling took into account the number of endocrinologists, patient accessibility to specialty diabetes care, and managed care in the geographic locations.

The intent-to-treat (ITT) population was defined as all randomized subjects who received at least one injection of randomized medication starting from the evening of day 1. All efficacy and safety analyses were performed on the ITT population, with the exception of the percentage of subjects achieving HbA_1c ≤7% by week 30. For the latter analysis, the more clinically relevant population of evaluable subjects was used (see below). For ITT subjects who had recorded values for at least one scheduled visit subsequent to the baseline measurement, missing data (including missing values at intermediate visits) were imputed from scheduled visits using the last observation carried forward method. The least square means and SEs were derived from the general linear model for each treatment. Pairwise comparisons of the treatment effects were performed using Fisher’s protected testing procedure to control type I errors due to multiple comparisons (20). Similar analyses were performed for body weight, each fasting metabolic parameter, and postprandial plasma glucose concentrations without adjusting for the multiple comparison. Results are given as means ± SE unless otherwise indicated

The evaluable population was defined as all randomized subjects who completed treatment through week 30 and received at least 80% of the study medication injections. Subjects who missed 7 consecutive days of injections during the last 2 months of the study were excluded.

All safety analyses were performed using the ITT population. Treatment-emergent adverse events were defined as those occurring upon or after receiving the first randomized dose. The intensity of hypoglycemic episodes was defined as mild/moderate or severe. For mild/moderate hypoglycemia, subjects reported symptoms consistent with hypoglycemia that may have been documented by a plasma glucose concentration value (<60 mg/dl). For severe hypoglycemia, subjects required the assistance of another person to obtain treatment for their hypoglycemia, including intravenous glucose or intramuscular glucagon.

Plasma analytes were quantitated by Quintiles Laboratories (Smyrna, GA) or Esoterix Endocrinology (Calabasas Hills, CA) using standard methods. Serum insulin was quantitated by a two-site sandwich chemiluminescent immunoassay, and serum proinsulin was quantitated by a two-site immunochemiluminometric assay. HbA_1c was measured using a high-performance liquid chromatography methodology (21,22). Plasma exenatide and anti-exenatide antibodies were measured as described previously (8).

Three hundred seventy-seven subjects were randomized to treatment and received at least one dose of study medication (ITT population), 260 subjects completed the entire study (69%), and 117 withdrew early (31%) (Fig. 1). All subjects were treated with a sulfonylurea (45% glipizide, 33% glyburide, 20% glimepiride, 1% tolazamide, and 0.3% chlorpropamide). Thirty-nine percent of ITT subjects were also treated with an ACE inhibitor, 34% with an antithrombotic agent, and 37% with a serum lipid-reducing agent.

HbA_1c values declined in all treatment arms during the period between screening and randomization, averaged 8.6% at baseline, and were comparable across treatment arms (Fig. 2A). HbA_1c values declined in both exenatide arms during the initial 12 weeks of the study, in contrast to relatively little change in the placebo arm. Thereafter, HbA_1c values in the exenatide arms plateaued, followed by a slight rise toward baseline by the end of the study in parallel with a similar change in the placebo arm. At week 30, the HbA_1c change from baseline was −0.86 ± 0.11% in the 10-μg exenatide arm and −0.46 ± 0.12% in the 5-μg exenatide arm compared with an increase of 0.12 ± 0.09% in the placebo arm (adjusted P ≤ 0.0002 for pairwise comparisons). For the ITT population at week 30 with baseline HbA_1c >7% (n = 353), 41 subjects (34.2%) in the 10-μg exenatide arm and 31 subjects (26.7%) in the 5-μg exenatide arm reached an HbA_1c ≤7%, and these proportions of the population were significantly greater than in the placebo arm (9 subjects [7.7%]; P < 0.0001 for pairwise comparisons). For the evaluable population at week 30 with baseline HbA_1c >7% (n = 237), 33 subjects (41.3%) in the 10-μg exenatide arm and 28 subjects (32.6%) in the 5-μg exenatide arm reached an HbA_1c ≤7%, and these proportions of the evaluable population were significantly greater than in the placebo arm (6 subjects [8.8%]; P ≤ 0.0002 for pairwise comparisons).

When stratified by baseline HbA_1c ≥9%, the 10- and 5-μg exenatide arms had changes in HbA_1c from baseline of −1.22 ± 0.19% (n = 46) and −0.58 ± 0.24% (n = 46), respectively, compared with an increase of 0.13 ± 0.17% in the placebo arm at week 30 (n = 46; adjusted P < 0.05 for pairwise comparisons) (Fig. 2B). For subjects with baseline HbA_1c <9%, the 10- and 5-μg exenatide arms had changes in HbA_1c from baseline of −0.65 ± 0.12% (n = 83) and −0.39 ± 0.12% (n = 79), respectively, compared with an increase of 0.11 ± 0.12% in the placebo arm at week 30 (n = 77; adjusted P < 0.01 for pairwise comparisons).

Baseline fasting plasma glucose concentrations were similar across treatment arms (Fig. 1). By week 30, fasting plasma glucose concentrations in the 10- and 5-μg exenatide arms were reduced by −0.6 ± 0.3 and −0.3 ± 0.2 mmol/l from baseline, respectively, compared with an increase of 0.4 ± 0.3 mmol/l in the placebo arm (P < 0.05 vs. placebo for the 10-μg arm only).

Body weights averaged ∼96 kg at baseline (Fig. 1) and were slightly higher in the placebo arm than in the exenatide arms. Subjects in the 10-μg exenatide arm had progressive weight loss over the entire 30 weeks, with an end-of-study loss of −1.6 ± 0.3 kg from baseline (P < 0.05 vs. placebo) (Fig. 3). Subjects in the 5-μg exenatide arm had an end-of-study weight loss of −0.9 ± 0.3 kg from baseline (NS vs. placebo), and subjects in the placebo arm had an end-of-study weight loss of −0.6 ± 0.3 kg from baseline.

Baseline fasting insulin and proinsulin concentrations were similar across treatment arms (Fig. 1), and there were no significant differences in fasting plasma insulin concentrations across treatment arms over the course of the study. However, there was a significant reduction in fasting proinsulin concentrations in the 10-μg exenatide arm compared with baseline (−16 pmol/l, 95% CI −26.1 to −6.0) and with placebo (P < 0.01), with a similar trend noted in the 5-μg exenatide arm. Overall, there was a dose-dependent decrease in the proinsulin-to-insulin ratio toward more physiological proportions. Baseline proinsulin-to-insulin ratios were 0.66 ± 0.04, 0.59 ± 0.03, and 0.64 ± 0.04 in the 10-μg exenatide, 5-μg exenatide, and placebo arms, respectively. In the 10-μg exenatide arm at week 30, the mean proinsulin-to-insulin ratio was reduced −0.13 compared with baseline and was significantly lower than that in placebo (P = 0.001). There was a similar trend in the 5-μg exenatide arm.

There were no adverse trends apparent in vital sign measurements, physical examination findings, heart rate, or blood pressure between the treatment arms. Twelve subjects had mild-to-moderate abnormalities in their blood creatine phosphokinase concentrations; however, all changes were transient, with no consistent pattern. There were small reductions in LDL (P < 0.05 for pairwise comparisons) and apolipoprotein B (P < 0.05 for pairwise comparisons) concentrations in exenatide arms compared with placebo. However, other lipid parameters (total cholesterol, triglycerides, LDL-to-HDL ratios) did not differ significantly among treatment arms.

The incidence of serious treatment-emergent adverse events was low, with no discernable treatment pattern (4% in the 10-μg exenatide arm, 3% in the 5-μg exenatide arm, and 8% in the placebo arm). One subject in the 10-μg arm and one subject in the placebo arm experienced a myocardial infarction, and one subject in the placebo arm experienced clinical manifestations of coronary artery disease.

The most frequent adverse events were generally mild or moderate in intensity and gastrointestinal in nature (Table 1). The incidence of treatment-emergent nausea was generally mild or moderate in intensity and peaked during the initial weeks of dosing (weeks 0–8), then decreased in incidence thereafter (Fig. 4). In contrast, weight loss was progressive over 30 weeks, and subjects who never reported nausea also lost weight (Table 2). The incidence of severe nausea was low (5% in the 10-μg exenatide arm, 6% in the 5-μg exenatide arm, and 2% in placebo arm). Withdrawals due to nausea were also low (4% in the 10-μg exenatide arm, 2% in the 5-μg exenatide arm, and 0% in placebo arm).

There were no cases of severe hypoglycemia. The overall incidence of mild-to-moderate hypoglycemia was 36% in the 10-μg exenatide arm, 14% in the 5-μg exenatide arm, and 3% in the placebo arm. Only one subject withdrew due to hypoglycemia (5-μg exenatide arm). The incidence of treatment-emergent, dose-dependent hypoglycemia peaked during the initial weeks of dosing, then decreased over time.

Subjects who had a detectable anti-exenatide antibody titer at any time during the study were considered antibody positive for the purpose of stratifying treatment-emergent adverse events. The presence of anti-exenatide antibodies (41% at 30 weeks) had no predictive effect on glycemic control or adverse events. Most subjects with treatment-emergent anti-exenatide antibodies developed low titer antibodies of unknown biological relevance.

In most individuals with type 2 diabetes, hyperglycemia results from a failure of β-cell insulin secretory capacity to adequately compensate for insulin resistance in peripheral tissues (23). Results from the U.K. Prospective Diabetes Study (4–6) indicate that β-cell failure is progressive despite therapy with diet, metformin, sulfonylureas, or insulin. Reductions in HbA_1c have been shown to lower the risk of microvascular complications. Unfortunately, glycemic control in this population is often inadequate, with average HbA_1c values well above 8% (4–6,24). In addition, many available therapeutic treatments may not be tolerated by all patients or may have undesirable side effects, such as weight gain, hypoglycemia, and edema, that can impede the attainment of glycemic control and discourage patient compliance (1–6,25).

Exenatide is an incretin mimetic, having glucoregulatory activities similar to those of mammalian hormone GLP-1. These actions include glucose-dependent enhancement of insulin secretion, suppression of inappropriately high glucagon secretion, and slowing of gastric emptying (7–12,26). Exenatide’s glucose-dependent enhancement of insulin secretion may be mediated by exenatide binding to the pancreatic GLP-1 receptor (27). In animal models of diabetes and in insulin-secretory cell lines, exenatide and GLP-1 reportedly improve β-cell function by increasing the expression of key genes involved in insulin secretion, by increasing insulin biosynthesis, and by augmenting β-cell mass through multiple mechanisms (9,28). Data obtained in animal models (9,10,28,29) also indicate that exenatide and GLP-1 reduce food intake, cause weight loss, and have an insulin-sensitizing effect.

The data from the current trial demonstrate that long-term use of exenatide at fixed doses of 5 and 10 μg b.i.d. dose dependently improve overall glycemia (HbA_1c) in patients failing sulfonylurea therapy. Previous studies have documented how other therapies (i.e., acarbose, metformin, or thiazolidinediones), when added to a background of sulfonylurea, also elicit a glucose-lowering effect. It is difficult to draw comparisons with this current data because the majority of such studies (30–35) have observed drug effects in patients with worse glycemic control, hence much higher baseline HbA_1c levels, where HbA_1c lowering occurs more readily with intervention. A recent study (36) that approximates the treatment conditions of our study showed that when either metformin or pioglitazone are added to background sulfonylurea therapy, one observes HbA_1c lowering similar to that observed in our study. As each therapeutic agent comes with its own benefits and potential tolerability issues and safety considerations, further studies are necessary to better understand the available therapeutic options when choosing adjunctive therapy for sulfonylureas.

When assessing the glycemic effect of a therapeutic, it is perhaps more important to ascertain the proportion of subjects who achieved a goal of HbA_1c ≤7%. At the 10- and 5-μg doses, 41 and 33% of subjects who completed the study achieved this goal, respectively, compared with 9% of subjects administered placebo. In addition, more subjects withdrew from the study due to loss of glycemic control in the placebo arm than in the exenatide arms. The exenatide glucose-lowering effect would appear to be attributed to a robust effect on daytime postprandial glycemia (7,8) because the glucose-lowering effect on fasting plasma glucose was modest in both exenatide treatment arms (a significant reduction in the 10-μg arm but not in the 5-μg arm). The reduction (∼−0.9% at week 30) was in line with reductions in HbA_1c reported in a 28-day phase 2 study of −0.7 to −1.1% (8); however, due to differences in study designs, inclusion criteria, and dosing regimens, it is difficult to directly compare these results. The proinsulin-to-insulin ratio improved over the 30 weeks of exenatide treatment, suggesting an improvement in β-cell function (37).

Exenatide treatment at a fixed dose of 10 μg was associated with reductions in body weight that did not appear to plateau by week 30. This progressive weight loss is consistent with the known ability of exenatide to reduce food intake (12,29). Nausea may be suspected to be the cause for the weight loss; however, the incidence of nausea was greatest in the first few weeks following initiation of therapy, whereas weight loss was progressive over 30 weeks. Moreover, subjects who never reported nausea also lost weight, thus emphasizing the dissociation of the two effects. Although the effect on plasma lipids was not a primary objective of the study, the effect of exenatide to improve overall glycemia while causing weight loss, most significantly in the 10-μg arm, leads one to anticipate potential effects on circulating lipids. That said, there was a small reduction in LDL cholesterol and apolipoprotein B levels, but other lipid parameters were unchanged.

Overall, exenatide was generally well tolerated. There did not appear to be any evidence of a clear safety concern, but there were some important observations pertaining to tolerability in some patients. The most common treatment-emergent adverse event was dose-dependent nausea, and this was most notable at the time of initiating therapy and was reported at lower incidence thereafter. Nausea was mostly mild or moderate in intensity, with a low incidence of severe nausea (∼6%) and low withdrawal from the study due to nausea (∼3%).

Hypoglycemia occurred more readily in the exenatide-treated patients in a dose-dependent fashion. Events were of mild or moderate intensity with no severe hypoglycemia reported (requiring the assistance of another person). Interestingly, exenatide itself does not appear to intrinsically increase the risk of hypoglycemia because in a similarly designed study, when it was added to a background of metformin therapy, there was no increase in reported hypoglycemia, even though overall glycemia had improved (38). With this in mind, it is possible to speculate that the increase in hypoglycemia observed in this study was likely caused by the background susceptibility to hypoglycemia often observed in sulfonylurea-treated patients coupled with lower ambient glycemia.

In summary, exenatide reduced HbA_1c and was associated with sustained weight loss. The most frequent adverse events were mild or moderate and gastrointestinal in nature. The incidence of hypoglycemic risk associated with sulfonylurea treatment increased with exenatide administration as overall glycemic control improved. Long-term use of exenatide at fixed subcutaneous doses of 5 and 10 μg b.i.d. appears to have potential for the treatment of patients with type 2 diabetes not adequately controlled with sulfonylurea agents, with 41% able to reach and maintain an HbA_1c ≤7% in the 10-μg b.i.d. arm at the end of 30 weeks.

Ahmann A, Albery R, Albu J, Angelo J, Argoud G, Banov C, Baron M, Beasey M, Black J, Blonde L, Bock A, Bradley V, Buse J, Canadas R, Casner P, Cathcart H, Cohen L, Collins G, Conway M, Corder C, Cyrus J, Daboul N, de la Garza C, DeFronzo R, Duckor S, Durden J, Eliosoff R, Farnsworth K, Farrell J, Fishman N, Freedman L, Gaman W, Gavin L, Gee D, Goldstein B, Hamad M, Harvey W, Havlicek R, Herring C, Heuer M, Holloway R, Horowitz B, James D, Jost-Vu E, Kaplan R, Kawley A, Kennedy L, Kim K, Klaff L, Klein E, Klonoff D, Levinson L, Littlejohn T, McIlwain H, McInroy R, Miller J, Miller S, Mills R, Moretto T, Mudaliar S, Myers L, Norwood P, Osei K, Patel M, Patel N, Pullman J, Raad G, Radparvar A, Riff D, Rigby S, Robinson J, Rood R, Rosenstock P, Saponaro J, Schmidt L, Shockey G, Schumacher D, Schwartz S, Shapiro J, Shapiro W, Snyder J, Sugimoto D, Sullivan J, Taber L, Troupin B, Ward W, Warren K, Weinstein R, Weiss D, Weiss R, Weissman P, Whitehouse F, Williams K, Wofford M, Wright D, Wysham C, Yates S, Zayed A, Zegarelli L, Zigrang W.

This trial was supported by Amylin Pharmaceuticals and Eli Lilly.

The authors thank the Exenatide-113 Clinical Study Group for their excellent assistance in the conduct, reporting, and quality control of the study and all of the patients who volunteered to participate. In addition, the following individuals are gratefully acknowledged for their valuable contributions to the conduct, reporting, and quality control of the study and to the development of the manuscript: Miriam Ahern, Roberta Allen, Maria Aisporna, Thomas Bicsak, Amy Halseth, John Holcombe, Orville Kolterman, Szecho Lin, David Maggs, Loretta Nielsen, Terri Poon, Larry Shen, Michael Sierzega, Kristin Taylor, Michael Trautmann, Amanda Varns, Barbara Wilkinson, Matthew Wintle, and Liping Xie.