New developments in incretin-based therapies: The current state of the field


  • Carolyn Robertson APRN, MSN, CDE

    Diabetes Nurse Educator, Corresponding author
      Carolyn Robertson, APRN, MSN, CDE, c/o Allen Kaufman, MD, 1555 Third Avenue, New York, NY 10129.
      Tel: 516-384-0564;
      Fax: 646-843-4764;
    Search for more papers by this author

Carolyn Robertson, APRN, MSN, CDE, c/o Allen Kaufman, MD, 1555 Third Avenue, New York, NY 10129.
Tel: 516-384-0564;
Fax: 646-843-4764;


Purpose: To update readers on developments in incretin therapies since the previous JAANP supplement in 2007; specifically, to describe clinical data for currently available incretin-based therapies as well as those under consideration by regulatory agencies.

Data source: Medline search for peer-reviewed publications.

Conclusions: Incretin-based therapies have pharmacologic properties that avoid some key limitations of previous treatments, such as hypoglycemia and weight gain. Certain agents also lower blood pressure and have the potential to reduce cardiovascular risk. The insulin-secreting action of incretin-based therapies only occurs under hyperglycemic conditions, thus minimizing the risk of hypoglycemia, unless combined with a sulfonylurea. The DPP-4 inhibitors are orally administered and demonstrate modest A1c reductions (0.6%–0.8%); the best results occur when combined with metformin. Glucagon-like peptide-1 (GLP-1) receptor agonists liraglutide and exenatide have shown greater A1c reductions (typically ≥ 1.1% and as high as 1.7%), and these agents have beneficial ancillary effects, including weight and systolic blood pressure reduction. Both DPP-4 inhibitors and GLP-1 receptor agonists have shown the ability to improve pancreatic beta-cell function in early studies.

Implications for practice: Data are provided on the efficacy and tolerability of approved incretin therapies, and on treatments currently in regulatory review, in order to inform readers and guide their practice.


Financial sponsorship for development of this article was provided by Novo Nordisk Inc., Princeton, NJ. The author has been paid by Amylin Pharmaceuticals and Eli Lily and Company for lectures and continuing education presentations on this topic; however, no monetary or other inducement was made to submit this article. The author wishes to thank Esther Nathanson, Helen Marshall, and Beverly LaFerla at Watermeadow Medical for their editorial assistance.


The incretin-based therapies are emerging as effective treatments for type 2 diabetes mellitus (T2DM) (Holst, Vilsbøll, & Deacon, 2009). Moreover, these agents have pharmacologic properties that avoid some key limitations of previous treatments, such as hypoglycemia and weight gain. Certain agents also lower blood pressure and, therefore, have the potential to reduce cardiovascular risk. Preliminary human data suggest an ability to enhance beta-cell function, whereas animal studies have shown improved beta-cell mass and therefore the potential to delay disease progression. As a result, key diabetes associations, including the American Diabetes Association (ADA), the American Association of Clinical Endocrinologists (AACE), and the European Association for the Study of Diabetes (EASD), consider initiation of certain incretin agents appropriate as a second-tier therapy in patients whose A1c remains above 7% despite lifestyle modification and use of metformin. The associations have acknowledged this with the inclusion of glucagon-like peptide-1 (GLP-1) agonists in recent guidelines (Jellinger et al., 2007; Nathan et al., 2008). As the volume of clinical trial data on incretin-based therapies increases and long-term safety data become available, nurse practitioners need to understand and be able to differentiate between the available drugs; they must also retain familiarity with the treatment choices available, and be able to explain to patients the clinical outcomes and tolerability/safety issues they can expect.

A previous supplement to this journal, published in June 2007 (JAANP, 2007), summarized early data on the first GLP-1 receptor agonist, exenatide, and the DPP-4 inhibitor sitagliptin. Since then, phase 3 trials on new drugs, including the human GLP-1 analog liraglutide and DPP-4 inhibitors alogliptin and saxagliptin, have been completed. In addition, some longer-term studies of exenatide have been reported, while a long-acting release (LAR) formulation of exenatide, several DPP-4 inhibitors, and another human GLP-1 analog have entered into phase 3 trials.

With the rising cost of health care, concerns about the price of novel therapies naturally arise. Although the upfront cost of incretin agents is typically somewhat greater than that of traditional blood glucose-lowering agents (e.g., monthly costs for exenatide twice daily, based on wholesale costs, have been shown to be $155–$220, compared with $115–$165 for twice-daily rosiglitazone) (Bond, 2006; Ezzo & Ambizas, 2006), their ability to help patients avoid hypoglycemia as well as the potential to limit disease progression, including possibly the need for insulin, may swing the balance in their favor. In the case of GLP-1 agonists, the additional benefit of addressing cardiovascular disease risk factors may be particularly attractive from a cost-effectiveness perspective (Minshall et al., 2008).

This article describes clinical data for the incretin-based therapies that are currently available as well as those under consideration by regulatory agencies.

The state of knowledge in 2007

In 2007, two incretin therapies were available, with most clinical experience gained for exenatide (Byetta, Amylin Pharmaceuticals Inc., San Diego, CA), which was launched in 2005. The 2007 JAANP supplement on incretin therapies (JAANP, 2007) summarized the results of three phase 3, 30-week studies investigating the addition of exenatide in patients with T2DM poorly controlled on either metformin (DeFronzo et al., 2005), a sulfonylurea (SU) (Buse et al., 2004), or both (Kendall et al., 2005). In addition, this publication reported data from a 1-year open-label extension of the above (Blonde et al., 2006), and a 16-week trial of exenatide plus a thiazolinedione (TZD) with or without metformin (Zinman et al., 2007).

In October 2006, the first DPP-4 inhibitor, sitagliptin (Januvia, Merck & Co., Inc, Whitehouse Station, NJ), was launched in the United States and was available worldwide by April 2007. Sitagliptin is approved at a dose of up to 100 mg/day, either as monotherapy or initial therapy with metformin, or as add-on combination therapy with either pioglitazone or glimepiride with or without metformin (Januvia Prescribing Information, 2009). Following approval in the European Union, vildagliptin (Galvus, Novartis Pharmaceuticals, East Hanover, NJ) was filed with the Food and Drug Administration (FDA) in 2007, but its release in the United States has since been suspended indefinitely because of dermatologic side effects in animal studies. Reported placebo-corrected A1c reductions attainable with DPP-4 inhibition were in the range of 0.6–0.8%, without weight change, during monotherapy (Aschner et al., 2006; Charbonnel et al., 2006; Raz et al., 2006). Sitagliptin is well tolerated, with a low incidence of gastrointestinal (GI) side effects and no documented incidence of hypoglycemia. This article will not discuss vildagliptin in detail because of its limited use in the United States but, to summarize briefly, when dosed as recommended (100 mg once daily or 50 mg twice daily, except when combined with an SU, when the recommended dose is 50 mg once daily) (EMEA 2007 Scientific Discussion on Vildagliptin, 2007), vildagliptin reduces A1c by approximately 0.9% from baseline, and up to 1.1%, as monotherapy (Pi-Sunyer, Schweizer, Mills, & Dejager, 2007; Rosenstock, Baron, Dejager, Mills, & Schweizer, 2007), or in combination with metformin (Bosi, Camisasca, Collober, Rochotte, & Garber, 2007) or a TZD (Garber, Schweizer, Baron, Rochotte, & Dejager, 2007). Fasting plasma glucose (FPG) reductions are in the range 20–30 mg/dL, and the effects of vildagliptin on weight are negligible. Adverse effects are similar to those observed with other DPP-4 inhibitors (Richter, Bandeira-Echtler, Bergerhoff, & Lerch, 2008).

Update on exenatide

Since publication of the previous JAANP supplement, a long-term pooled analysis of data from 283 patients from the original combined AMIGO study cohort (n = 974) who continued exenatide beyond 82 weeks has been published. In this completer group, the A1c improvement seen at 30 weeks was maintained after 2 years of exenatide treatment (p < .05 vs. baseline) (Figure 1), with additional reductions in FPG (−25.2 mg/dL) and continued weight loss (4.7 kg; p < .001 vs. baseline) (Buse et al., 2007). There was also evidence of reduced systolic (−2.6 mmHg; p = 0.003) and diastolic (−1.9 mmHg; p < .001) blood pressure. Beta-cell function, measured using homeostasis model assessment-B (HOMA-B) levels, also improved significantly in the subgroup (n = 112) whose levels were measured. These data are encouraging because a deterioration of control over this time might have been anticipated with traditional antidiabetic drugs. Furthermore, in an extension study of 217 obese patients (mean body mass index = 34 kg/m2) from the original AMIGO cohort, the reduction in A1c after 12 weeks (−1.1%) was sustained at 3 years (−1.0%; p < .0001 vs. baseline) in subjects receiving exenatide plus metformin with or without SUs (Klonoff et al., 2008). Almost half of this cohort reached A1c ≤ 7%, and multiple markers of liver function and HOMA-B also improved. For readers interested in understanding the HOMA test, it estimates steady-state beta-cell function and insulin sensitivity as percentages of a normal reference population. The results correspond, but are not equivalent, to nonsteady state estimates derived from models such as hyperinsulinemic and hyperglycemic clamps, and intravenous and oral glucose tolerance tests (Oxford Centre for Diabetes, Endocrinology and Metabolism, 2007). The progressive reduction in body weight and improved lipid parameters in the trial above were documented in the previous JAANP supplement. As weight loss maintained over 1 year has been shown to improve cardiovascular disease risk factors including lipids and blood pressure (Look AHEAD Research Group, 2007), these data suggest the potential for cardiovascular disease risk reduction. However, because not all patients who started the AMIGO trials continued throughout the extension studies, these data may not be fully indicative of the clinical experiences we can expect.

Figure 1.

An A1c reduction of 1.1% in patients taking exenatide for 30 weeks was maintained when treatment was continued for 2 years. Reprinted from Buse et al. (2007) with permission from Excerpta Medica, Inc.

In clinical studies, exenatide has been compared with the insulin analogs biphasic insulin aspart 70/30 (Malone et al., 2008; Nauck et al., 2007a, 2007b) and insulin glargine (Heine et al., 2005). In both comparisons, exenatide achieved similar A1c reduction with superior postprandial glucose (PPG) control, which may be of clinical importance because postprandial hyperglycemia is known to be an independent risk factor for cardiovascular events (Fava, 2008). Although these results appear impressive, it must be remembered that the blood glucose-lowering potency of insulin is dose-dependent, so the efficacy advantages of exenatide would probably have been lost had higher insulin doses been given. However, higher insulin doses would have increased the risk of hypoglycemia as well as weight gain, both of which are very low with the incretin-based therapies.

A long-acting exenatide formulation, exenatide LAR, is currently being evaluated for once-weekly (QW) dosing. In a 15-week placebo-controlled phase 2 trial in 45 patients, exenatide, 0.8 or 2.0 mg given weekly, reduced mean A1c by 1.4% and 1.7%, respectively (p < .0001 vs. placebo, both doses) (Kim et al., 2007), allowing 36% and 86% of patients to achieve A1c ≤ 7%. FPG decreased by approximately 40 mg/dL in both groups (p < .001 vs. placebo), and body weight fell by a mean of 3.8 kg in patients receiving the higher dose (p < .05). Injection-site bruising has been noted during exenatide LAR trials and, in this study, occurred more frequently with exenatide LAR than placebo (13% and 7% in lower and higher dose groups vs. 0% of placebo patients). Of note is that injections were administered by study site staff; when patients start self-injecting, it is likely that such bruising will occur more frequently than described above, at least until patients become familiar with the optimal injection technique. Nausea was the most common adverse reaction, occurring in 19% and 27% of patients receiving the lower and higher doses, respectively (vs. 15% with placebo); gastroenteritis occurred in 19% and 13% of patients in the lower and higher dose groups, respectively. Mild hypoglycemia occurred in 25% of patients, in the lower-dose LAR group only.

An open-label 30-week phase 3 study of 295 patients previously managed with diet/exercise and/or oral antidiabetic drugs (OADs) has recently compared exenatide LAR with twice-daily exenatide (Drucker et al., 2008b). Here, 2 mg of the LAR formulation given QW achieved a greater reduction in A1c (−1.9%) than exenatide 10 μg twice daily (−1.5%, p = .002) as well as greater FPG reductions (42 vs. 25 mg/dL; p < .0001). Weight loss of about 4 kg occurred with both formulations, with no major hypoglycemia. Significant improvement in several quality of life measures was observed with both treatments, but the reduced dosing frequency afforded by the QW formulation was, perhaps surprisingly, not associated with greater treatment satisfaction (Best et al., 2008; Kim, Boye, Wintle, Hufford, & Rubin, 2008).

The most serious tolerability concern raised to date in association with twice-daily exenatide is acute pancreatitis. This concern has arisen as a result of 30 individual case reports (U.S. FDA, 2007), which, in 2008, prompted the FDA to issue a safety notice that suggested antidiabetic therapies other than exenatide be considered in patients with a history of pancreatitis (U.S. Food and Drug Administration [FDA], 2008). Although the available evidence suggests that the incidence of pancreatitis in patients receiving exenatide may be no higher than that of the general population (Frey, Zhou, Harvey, & White, 2006; Nathan et al., 2008), the revised label for exenatide nevertheless advises close monitoring for symptoms associated or consistent with pancreatitis, discontinuation in patients whose symptoms are consistent with pancreatitis until the cause is identified, and termination of treatment where no other cause can be found.

Liraglutide: The first human GLP-1 receptor analog

A significant update since the last JAANP review is the development of the once-daily human GLP-1 receptor agonist liraglutide (Novo Nordisk A/S, Bagsværd, Denmark) which is currently under regulatory review. Liraglutide shares 97% similarity in structure (homology) with native GLP-1; however, since the native GLP-1 hormone has a very short half-life (approximately 2 min), making it unsuitable for therapeutic use, structural modifications have been made to extend the half-life of liraglutide to approximately 13 h, making it suitable for once-daily dosing (Agersø, Jensen, Elbrønd, Rolan, & Zdravkovic, 2002).

The main body of phase 3 clinical trial data on liraglutide comes from the Liraglutide Effect and Action in Diabetes (LEAD) studies, a series of phase 3 clinical trials investigating its use at different stages of diabetes progression. Liraglutide was started at 0.6 mg daily, with weekly incremental increases of 0.6 mg until the allocated daily dose of 1.2 or 1.8 mg was reached. In the 1-year study of liraglutide monotherapy in 745 patients previously treated with either diet and exercise only, or OAD monotherapy, mean A1c in the lower and higher dose groups decreased by 0.84% (p = .001 vs. glimepiride) and 1.14% (p < .0001 vs. glimepiride), respectively, compared with a reduction of 0.51% with glimepiride monotherapy (Garber et al., 2009). The percentage of patients reaching the ADA target A1c of < 7.0% was 43% (p = .0007) and 51% (p < .0001) vs. 28% with glimepiride. Data showing the effects of liraglutide on A1c for all LEAD studies are presented in Figure 2. Respective decreases in FPG values were 15.1 and 25.6 mg/dL with liraglutide, compared with 5.2 mg/dL for glimepiride. Participants in the liraglutide group lost 2.1–2.5 kg in weight, whereas glimepiride recipients gained 1.1 kg (p < .0001 vs. both).

Figure 2.

(A) Effect of liraglutide on A1c: data across all LEAD studies. Data from Garber et al. (2009), Marre et al. (2009), Nauck et al. (2009a), Russell-Jones et al. (2009), and Zinman et al. (2009). (B) Effects on A1c of drugs other than liraglutide that are under regulatory review. Data are from Allen et al. (2008), Chen et al. (2008), DeFronzo et al. (2008a), Drucker et al. (2008), Nauck et al. (2008), Pratley et al. (2008b), Ravichandran et al. (2008), and Rosenstock et al. (2008a, 2008c). *Significant vs. comparator; #Change in HbA1c from baseline for overall population (LEAD-4,-5) add-on to diet and exercise failure (LEAD-3); or add-on to previous OAD monotherapy (LEAD-2,-1).

The remaining LEAD studies assessed liraglutide in combination with one or more OADs. In a 26-week study of 1091 patients that compared liraglutide 1.2 and 1.8 mg daily with glimepiride or placebo, each added to metformin, both doses of liraglutide and glimepiride lowered A1c by approximately 1.0% (p < .0001 for all groups vs. placebo) (Nauck et al., 2009a). Liraglutide was also associated with improvements in beta-cell function, as evidenced by a 28% increase in HOMA-B levels; similar improvement was observed with glimepiride. Weight loss of approximately 2.7 kg was observed in liraglutide-treated patients compared with weight gain of 1.0 kg in those receiving glimepiride (p < .0001). Systolic blood pressure (SBP) decreased in liraglutide patients compared with those taking an SU (relative reduction in SBP = 3.2 [p = .01 vs. glimepiride] and 2.7 mmHg [p = .05 vs. glimepiride] for lower and higher liraglutide doses, respectively).

Liraglutide has also been compared with rosiglitazone as add-on therapy to glimepiride in a 26-week trial in 1018 patients (Marre et al., 2009). Liraglutide lowered A1c to a greater extent (−1.1% for both doses) than rosiglitazone (−0.4%; p < .0001) and, as in previous LEAD trials, improvements in SBP (approximately −2.7 mmHg) and beta-cell function were observed only with liraglutide. The negligible effects on weight in patients taking liraglutide plus glimepiride compared favorably to the gain of 2.1 kg in rosiglitazone plus glimepiride patients.

Liraglutide has also been studied in triple combination therapy (Zinman et al., 2009). In a placebo-controlled 26-week trial of 533 patients on metformin plus rosiglitazone, the addition of liraglutide, 1.2 or 1.8 mg daily, lowered A1c by 1.5% from baseline, with reduction in weight of 1.0 and 2.0 kg, respectively. There was also a relative reduction in SBP (6.7 and 5.6 mmHg, respectively, for lower and higher liraglutide doses). Another 26-week study, of 581 patients on metformin plus an SU, compared the addition of liraglutide 1.8 mg daily or insulin glargine (Russell-Jones et al., 2009). Here, liraglutide lowered A1c to a greater extent than glargine (−1.3% vs. −1.1%, p < .01), in association with a 1.8 kg reduction in weight which was notable when compared with the 1.6 kg weight gain in the insulin group. It should be noted, however, that insulin titration was not optimized.

Of particular interest is a recent 26-week comparison of the addition of either once-daily liraglutide (1.8 mg) or twice-daily exenatide (10 μg) to metformin and/or SU in 464 patients (Buse et al., 2009). In this trial, liraglutide showed a significantly greater reduction in A1c: (1.1% vs. −0.8% (p < .0001), enabling a greater proportion of patients to reach the ADA target of < 7%: 54% vs. 43%, respectively (p < .01). HOMA-B levels indicated greater improvement in beta-cell function with liraglutide (p < .0001), and weight loss was more marked in this group (−3.2 vs. −2.9 kg). Both treatments reduced SBP (−2.5 and −2.0 mmHg, respectively). Initially, nausea occurred in approximately 26% of patients in each group, but it resolved more quickly in liraglutide-treated patients (p < .0001). Two patients in the exenatide group who were also taking an SU experienced major hypoglycemia compared with none in the liraglutide group; minor hypoglycemia was also significantly less frequent with liraglutide (1.9 vs. 2.6 events/patient/year; p = .01).

Liraglutide has generally been well tolerated. Across the LEAD studies, there have been very low rates of minor hypoglycemia, ranging from 0.1 to 1.2 events/patient/ year, with only five major hypoglycemic events, all in patients taking SUs. This highlights the finding, also observed with exenatide treatment, that the risk of hypoglycemia increases when GLP-1 receptor agonists are combined with an SU. Also in common with exenatide, the most frequent side effects were (most commonly nausea), suggesting that this is a class effect with GLP-1 receptor agonists. In the 52-week study, nausea was reported in approximately 28% of patients, but was typically mild and transient (Garber et al., 2009). There was a low incidence of pancreatitis in the LEAD trials: 5 of 2420 patients representing a frequency lower than that seen in the general population (EMEA 2006 Scientific Discussion on Exenatide, 2006). Antibody formation to liraglutide has been minimal (Hermansen et al., 2008; Marre et al., 2009; Russell-Jones et al., 2009) and showed no correlation with efficacy. A key safety advantage for liraglutide compared with exenatide and sitagliptin is that its clearance does not depend on renal metabolism or excretion (Bjørnsdottir et al., 2008), thus avoiding the need for dose adjustment in patients with renal impairment.

A noteworthy finding in the LEAD studies (shared with exenatide) has been the apparent ability of liraglutide to improve certain cardiovascular risk factors that link with impaired glucose control to define the “metabolic syndrome.” At 1.8 mg, in particular, liraglutide has tended to decrease SBP and weight, and a recent meta-analysis based on data from 3967 patients across the six LEAD studies found that liraglutide 1.8 mg significantly reduced several cardiovascular risk biomarkers, including total, low-density lipoprotein and very low-density lipoprotein cholesterol, B-type natriuretic peptide (−12%), and high-sensitivity C-reactive protein (−23.1%) (p < .01 vs. placebo) (Plutzky, Garber, Falahati, Toft, & Poulter, 2009). These beneficial effects diminish when liraglutide is combined with an SU (Marre et al., 2009; Russell-Jones et al., 2009), probably reflecting the tendency of SUs as a group to induce weight gain and therefore negate some of the potential improvements in cardiovascular health.

One potential safety concern identified during the development program for liraglutide was an increase in C-cell hyperplasia and adenomas, identified during preclinical studies in rodents; at the highest dose of liraglutide, in some cases, C-cell carcinoma was also identified. In all cases, calcitonin, a biomarker of C-cell activation, was elevated both acutely and over a longer preceding time period. There were no such findings in nonhuman primates at very high doses. An extensive calcitonin screening program, including >5000 individuals, revealed no data to suggest that liraglutide activates human C-cells (McGill, 2009).

Weekly or 2-weekly dosed GLP-1 receptor agonists: Taspoglutide and albiglutide

Taspoglutide (Hoffmann-La Roche Ltd., Basel, Switzerland) is a long-acting human GLP-1 analog currently in phase 3 development and is being evaluated for QW or twice-weekly (Q2W) dosing. Results are now available from a study evaluating three taspoglutide doses (Kapitza et al., 2008), along with data from two phase 2 clinical trials (Nauck et al., 2009b; Ratner et al., 2008), all in obese patients inadequately controlled with metformin. These trials have shown that taspoglutide 20–40 mg can significantly reduce glucose exposure, lowering A1c by 1.0–1.2% from baseline (p < .0001 vs. placebo) (Nauck et al., 2009b; Ratner et al.), and reducing FPG (Kapitza et al.; Nauck et al.; Ratner et al.) as well as PPG excursions (Kapitza et al.; Nauck et al., 2009b). Results from QW dosing were generally better than those after Q2W dosing: 70–80% of patients receiving 10 or 20 mg QW reached A1c ≤ 7%, vs. 44–63% of patients using Q2W dosing. Weight loss was significant, progressive, and dose-dependent, ranging from 2.0 kg in the 10 mg QW and 20 mg Q2W groups to 2.8 kg in the 20 mg QW group. As with other GLP-1 agents, adverse effects were mainly GI, transient, and dose-dependent. Withdrawal rates from trials due to adverse events ranged from 6% to 12%; interestingly, dose escalation from 20 to 40 mg QW was not associated with improved glycemic control, nor with reduced tolerability (Ratner et al.). No serious drug-related adverse events, laboratory abnormalities, or injection-site reactions were reported. The only data reported for hypoglycemia indicate a very low frequency: seven events (five asymptomatic) in 6 of 306 patients, and none were severe (Nauck et al., 2009b).

The new recombinant GLP-1 receptor agonist albiglutide (GlaxoSmithKline, Brentford, London, U.K.) has been shown to reduce fasting and PPG levels in pharmacokinetic studies in T2DM; it entered phase 3 testing in early 2009 (Lovshin & Drucker, 2009).

New developments with DPP-4 inhibitors: Sitagliptin, sitagliptin/metformin, alogliptin, saxagliptin

The role of DPP-4 inhibitors in the management of T2DM is expected to be in combination with metformin in drug-naï ve patients and those inadequately controlled on metformin alone, or as monotherapy in patients for whom metformin is contraindicated or poorly tolerated (Ahren, 2008). Many similarities exist between the different DPP-4 inhibitors: all are absorbed rapidly, reaching peak concentrations within 1–2 h, and achieving almost complete inhibition of DPP-4 within 0.5 h (Ahren). A1c reductions generally fall in the range of 0.6%–1.1% with monotherapy, and the incidence of adverse events is typically no different from placebo; hypoglycemia is rare, although increased when DPP-4 inhibitors are combined with an SU (Ahren), possibly due to a mechanism by which SUs somehow disrupt the glucose dependency of GLP-1 (De Heer & Holst, 2007). Differences between agents relate primarily to their route of metabolism, such that sitagliptin is excreted almost entirely by renal mechanisms, therefore requiring dose adjustment in renal impairment (Bergman et al., 2007), whereas saxagliptin is metabolized to an active compound by the liver, after which both the drug and its metabolite are renally excreted. By contrast, alogliptin is minimally metabolized (Deacon, 2008). Hepatic impairment does not appear to affect the pharmacokinetics of the DPP-4 inhibitors, and there are no known drug interactions between these and any other drugs (Ahren).

Sitagliptin alone or in combination with metformin

Sitagliptin is the only DPP-4 inhibitor approved to date in the United States. In monotherapy (Goldstein et al., 2007, Raz et al., 2006), A1c reductions are modest at 0.6%–0.8% vs. placebo, and similar placebo-corrected A1c reductions have been documented when sitagliptin is added to TZDs (Rosenstock et al., 2006) or SUs (Hermansen et al., 2007). Greater A1c reductions have been possible when sitagliptin is combined with metformin: when added to existing metformin treatment, sitagliptin has achieved placebo-subtracted reductions in A1c of 1.0% (Raz et al., 2008) and, in initial combination with high-dose metformin (2000 mg), up to 2.1% (Goldstein et al.). These results were achieved without a significant effect on weight or incidence of hypoglycemia, and with only placebo-level adverse events (Figure 3). The combination of sitagliptin and metformin increased HOMA-B more significantly than did either agent as monotherapy (Goldstein et al.). Synergism between these two drug classes seems likely, because metformin stimulates GLP-1 secretion (Ahren, 2008), whereas sitagliptin increases the half-life of GLP-1. As such, one recent incretin-based agent combines sitagliptin 50 mg and metformin 500 or 1000 mg (Janumet, Merck & Co., Inc, Whitehouse Station, NJ). The combination therapy is taken twice daily with meals, with a starting dose based on the patient's current regimen; gradual dose escalation is advised to minimize GI disturbance (Janumet Prescribing Information, 2009). The most common adverse reactions are diarrhea (7.5%), upper respiratory tract infection (6.2%), and headache (5.9%).This sitagliptin/metformin combination is contraindicated in patients with hepatic disease and renal dysfunction (Janumet Prescribing Information, 2009); the warning regarding allergic and adverse hypersensitivity reactions observed with sitagliptin is retained in the combined preparation. The most serious, albeit rare, adverse reaction observed with the sitagliptin/metformin combination is lactic acidosis due to metformin accumulation, which has proven fatal in 50% of cases; regular monitoring of renal function is therefore necessary, and the drug should be immediately discontinued in suspected cases. Such potentially serious adverse effects underscore the need for close monitoring of creatinine levels and/or creatinine clearance in all patients receiving metformin, whether alone or in combination with other agents. Specifically, metformin should be avoided in any patient with renal disease or renal dysfunction (suggested by serum creatinine levels ≥1.5 mg/dL in men and ≥1.4 mg/dL in women or abnormal creatinine clearance) (FDA Glucophage prescribing information). Other adverse reactions to the combined sitagliptin/metformin product include rare but serious hypersensitivity reactions observed in patients treated with sitagliptin, including Stevens–Johnson syndrome and decreased vitamin B12 levels, resulting in the need for annual evaluation.

Figure 3.

Efficacy of sitagliptin used in monotherapy and when combined with metformin. Data are from Charbonnel et al. (2006), Goldstein et al. (2007), and Raz et al. (2008).

Recently, sitagliptin has been compared with exenatide, in a small 2 × 2-week cross-over trial (DeFronzo et al., 2008b). Although the study was too short to evaluate the effect on A1c, both drugs had similar effects on FPG, but 2-h PPG was lower with exenatide (133 vs. 208 mg/dL, p < .0001). Moreover, switching from sitagliptin to exenatide reduced PPG by 76 mg/dL, whereas switching from exenatide to sitagliptin increased it by 73 mg/dL. Exenatide also had a greater suppressive effect on glucagon secretion and gastric emptying, and reduced total calorie intake compared with sitagliptin (−134 vs. +130 kcal; p = .02).

All these observations are consistent with the findings of individual clinical trials, in which GLP-1 receptor agonists have tended to produce greater blood glucose-lowering effects than DPP-4 inhibitors. Moreover, the former have consistently decreased weight, whereas the DPP-4 inhibitors have been weight-neutral. These differences are probably explained by the fact that GLP-1 receptor agonists are dosed to produce pharmacologic concentrations, whereas the level of GLP-1 receptor stimulation that can be achieved with DPP-4 inhibitors is limited by endogenous GLP-1 secretion (Drucker & Nauck, 2006).

Although sitagliptin is generally well tolerated, rare but serious hypersensitivity and allergic reactions requiring discontinuation of treatment have been reported (anaphylaxis, angioedema, and exfoliative skin conditions including Stevens–Johnson syndrome) as well as a relative increase in infections compared with other DPP-4 inhibitors (Amori, Lau, & Pittas, 2007). It should also be noted that the standard 100 mg daily dose of sitagliptin should be reduced to 50 and 25 mg daily in patients with moderate or severe renal impairment, respectively (Januvia Prescribing Information, 2009).


Alogliptin (Takeda San Diego Inc., CA) is currently in regulatory review in the United States. Alogliptin monotherapy has been studied in drug-naï ve patients (DeFronzo, Fleck, Wilson, & Mekki, 2008a; Fleck, Ronald, Covington, Wilson, & Mekki, 2008), and in combination therapy with either glyburide (Pratley, Kipnes, Fleck, Wilson, & Mekki, 2008a), metformin (Nauck, Ellis, Fleck, Wilson, & Mekki, 2008), pioglitazone (with or without metformin or SU) (Pratley, Reusch, Fleck, Wilson, & Mekki, 2008b), or insulin (with or without metformin) (Rosenstock et al., 2008b). Initial data suggest that alogliptin, at doses of 12.5–100 mg/day, significantly improves glycemic control in T2DM patients inadequately controlled on prior treatments, without increasing weight or hypoglycemia; there appears to be no clinical benefit to doses exceeding 25 mg daily. In studies to date, maximum A1c reductions in patients taking the 25 mg dose were 0.8% after 26 weeks, when added to pioglitazone (Pratley et al., 2008b), and 0.7% when combined with insulin (Rosenstock et al., 2008b), with FPG reductions of 12–16 mg/dL (DeFronzo et al., 2008a; Rosenstock et al., 2008b). Reduction in PPG levels of approximately −35 mg/dL was also demonstrated in a 14-day single-dose study (Covington et al., 2008). Consistent with other DPP-4 inhibitors, alogliptin appears to be weight-neutral. In contrast to vildagliptin, there is no indication of dermatologic toxicity in animals (Sato, Ozaki, Salamon, Christopher, & Yamamoto, 2008).


Saxagliptin (Onglyza, Bristol-Myers Squibb Company, New York, NY), for which a new drug application was filed in June 2008, is another new DPP-4 inhibitor. In clinical trials of 12 and 24 weeks in drug-naï ve patients, saxagliptin monotherapy at doses of 2.5–10 mg showed significant reductions in A1c (0.6–1.7%), FPG (14–31 mg/dL), and PPG (24–106 mg/dL) with no effect on weight (Chen et al., 2008; Rosenstock et al., 2008a; Rosenstock, Sankoh, & List, 2008c); the higher dose was associated with greater glycemic benefits. In combination with metformin 500 mg, saxagliptin 5 or 10 mg lowered A1c by approximately 2.5% from baseline, with FPG and PPG reductions of approximately 60 and 138 mg/dL, respectively (Chen et al.). These results were all statistically significantly superior to either agent as monotherapy, as was the increase in HOMA-B levels. Improvements in glycemic control were also observed when saxagliptin was added to the treatment regimen in patients inadequately controlled on TZD monotherapy (Allen, Holander, Li, & Chen, 2008) and SU monotherapy (Ravichandran, Chacra, Tan, Apanovitch, & Chen, 2008). Dose adjustments due to hepatic impairment (Patel et al., 2008), age, or gender (Boulton, Goyal, Li, Kornhauser, & Frevert, 2008) are not necessary.

Additional DPP-4 inhibitors in late-stage development include linagliptin (Boehringer Ingelheim, Ingelheim, Germany) and dutogliptin (Phenomix, San Diego, CA).


The range of incretin-based therapies and the clinical trial database has grown substantially over the last couple of years, reflecting the growing popularity of these agents and their potential ability to overcome many of the limitations of conventional treatments. Most significantly, the insulin-secreting action of incretin-based therapies only occurs under hyperglycemic conditions, thus minimizing the risk of hypoglycemia, unless combined with an SU. Although DPP-4 inhibitors are orally administered, their efficacy is limited by endogenous GLP-1 secretion, and the best results seem to be obtained when these agents are combined with metformin. The GLP-1 receptor agonists liraglutide and exenatide have generally shown greater A1c reductions, and these agents have beneficial ancillary effects, including weight and SBP reduction. Both DPP-4 inhibitors and GLP-1 receptor agonists have shown the ability to improve beta-cell function in early studies.

The most significant recent developments are the release of phase 3 data for liraglutide and a long-acting formulation of exenatide that may enable weekly dosing, as well as emerging data on new GLP-1 receptor agonists with the potential for QW or Q2W dosing, and novel DPP-4 inhibitors with potential advantages over sitagliptin. In addition, we have an enlarged clinical database to document the utility of these agents as monotherapy and in combination with other treatments. For these reasons, the incretin therapies are now an established component of diabetes management.