Article Text

Randomised controlled trial
Adding subcutaneous liraglutide to metformin reduces HbA1c more than adding oral sitagliptin in patients whose type 2 diabetes is poorly controlled with metformin alone
  1. Vivian Fonseca1,
  2. Cyrus Desouza2,
  3. Amna N Khan1
  1. 1Tulane University Health Sciences Center, New Orleans, Louisiana, USA
  2. 2Omaha VA Medical Center, Omaha, Nebraska, USA
  1. Correspondence to Vivian Fonseca
    Tulane University Health Sciences Center, 1430 Tulane Ave, SL-53, New Orleans, LA 70112, USA; vfonseca{at}

Statistics from

Commentary on:


Incretin hormones have recently been recognised in playing an important role in the pathogenesis and treatment of type 2 diabetes.1 Of these, glucagon-like peptide (GLP-1) is probably the most important but has a limitation due to its rapid breakdown by the enzyme dipeptidyl peptidase 4 (DPP-4) giving it a half life in circulation of approximately 2 min.

To overcome this limitation, two pharmacological approaches have been developed: (1) DPP-4 inhibitors (sitagliptin, saxagliptin) that increase the endogenous concentrations of GLP-1 through inhibition of DPP-4 and (2) GLP-1 mimetics (exenantide) or analogues (liraglutide). The former have the advantage of being taken orally whereas the latter are injectable. Previous studies have all focused on comparing incretin-based treatments to other medications such as sulphonylureas, metformin, thiazolidinediones or insulin.2 3 There are very few studies comparing one form of incretin therapy to another.4 5


This was an active-comparator, parallel group, open-labelled study of 665 patients with type 2 diabetes conducted in several European countries, USA and Canada. Patients with HbA1c between 7.5% and 10% on metformin monotherapy (>1500 mg) were randomised to either metformin plus liraglutide 1.2 mg/day or liraglutide 1.8 mg/day or sitagliptin 100 mg/day. The primary outcome was change in HbA1c from baseline to 26 weeks. The effectiveness of liraglutide (vs sitagliptin) was assessed hierarchically by a non-inferiority comparison, with a margin of 0.4%, and then by a superiority comparison.


In this study, liraglutide 1.2 and 1.8 mg/day when added to metformin decreased HbA1c from baseline by −1.24% (95% CI −1.37 to −1.11) and −1.50% (95% CI −1.63 to −1.37), respectively. In contrast, sitagliptin reduced the A1c by −0.90% (95% CI −1.03 to −0.77). Patients with a higher baseline A1c had larger reductions in A1c. Weight loss was significantly greater for liraglutide than for sitagliptin specifically, −3.38 kg (95% CI −3.91 to −2.84) for 1.8 mg liraglutide; −2.86 kg (−3.39 to −2.32) for 1.2 mg liraglutide and −0.96 kg (−1.50 to −0.42) for sitagliptin. Nausea was more common with liraglutide than with sitagliptin.


With the development of many drugs for the treatment of diabetes, direct comparisons between drugs are important in order to help physicians make appropriate choices for their patients. The study by Pratley and colleagues is one such comparison and is therefore important. The results of the trial are not surprising and are similar to a previous trial comparing exenatide with sitagliptin.4 While it has the advantage of being a randomised study, unfortunately, since one drug was injectable and the other oral, a major limitation of this trial is that the study was open-label and not blinded (a double blind study using a double dummy approach should have been possible).

While both drugs raised GLP-1 levels, the injectable approach leads to a much higher circulating GLP-1 with pharmacological effects which would be expected to yield greater reductions in glucose through a variety of mechanisms. Second, because of a pharmacological effect on gastric emptying leading to increased satiety as well as possible effects on the satiety centre, GLP-1 analogues such as liraglutide leads to weight loss as well as delayed delivery of meal-related nutrients into the circulation. These effects are lacking with DPP-4 inhibitors such as sitagliptin, which only attempts to prolong circulating endogenous GLP-1 after a meal. On the other hand, DPP-4 inhibitors can raise other incretins such as gastric inhibitory polypeptide, although the clinical significance of such an effect is currently unknown and is probably negligible over the short duration of this study. Recent data points to a beneficial effect in the long term on α cell function with DPP-4 inhibition,3 6 whereas such an effect has been less well studied with GLP-1 analogues.

A longer-duration study comparing these two approaches to restore abnormalities in the incretin system in type 2 diabetes is needed. While the present study is well conducted and provides some useful immediate information, longer-term outcome data is needed before more rational choices can be made.


View Abstract


  • Competing interests VF has received consultancy fees, lecture fees and grants from Glaxo Smith Kline, Novartis, Novo-Nordisk, Takeda, Astra-Zeneca, Sanofi-Aventis, Eli Lilly, Daiichi-Sankyo, Novartis, NIH, ADA. CD has received consultancy fees from Novo Nordisk and Takeda. ANK is an investigator in the TECOS trial funded by Merck.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.