Impact of Liraglutide on Insulin Clearance in Patients with Type 2 Diabetes Mellitus: Evidence from a Randomized Controlled Trial
Article information
There is currently interest in the role of decreased insulin clearance in the pathophysiology and progression of type 2 diabetes mellitus (T2DM) [1,2]. Amidst this recognition of its potential clinical implications, an emerging question is the impact of diabetes medications on insulin clearance [1]. Notably, published reports of the impact of glucagon-like peptide-1 (GLP-1) receptor agonists are remarkably conflicted [3-9]. Indeed, in humans, endogenous GLP-1 has been associated with both increased hepatic insulin clearance and no effect thereupon [3-5]. Interventional studies have reported that GLP-1 infusion has no effect [6], while liraglutide increases [7], decreases [8] or does not change insulin clearance [9]. Surprisingly, there has been limited direct evaluation of this question in clinical trials. Our objective was thus to determine whether liraglutide impacts insulin clearance in patients with T2DM in a randomized, placebo-controlled trial.
The LIraglutide and Beta-cell RepAir (LIBRA) trial was a 48-week, double-blind, randomized, placebo-controlled trial designed to determine whether liraglutide can preserve β-cell function in early T2DM (clinicaltrials.gov: NCT01270789) [10]. The protocol and trial results have been described in detail [10]. In brief, patients with T2DM of modest duration underwent 4 weeks of intensive insulin therapy to ameliorate glucotoxicity-induced β-cell dysfunction, after which those with fasting glucose <7.0 mmol/L were randomized to daily subcutaneous injection of either liraglutide 1.8 mg or matching placebo for 48 weeks. The 51 randomized participants (mean age 58.2±8.1 years, body mass index [BMI] 30.2±5.0 kg/m2, prestudy A1c 6.8%±0.8%, duration of T2DM 2.6±1.9 years) underwent oral glucose tolerance tests (OGTT) at randomization and every 12-weeks thereafter, with study medication held on the morning of each OGTT. The protocol was approved by the Mount Sinai Hospital Research Ethics Board (REB no. 1182), and all participants provided written informed consent. As previously reported [10], the main outcomes of the trial showed that liraglutide yielded better β-cell function (primary outcome) and A1c over 48 weeks as compared to placebo. Liraglutide also lowered BMI compared to placebo, with no difference between the groups in insulin sensitivity, as measured by Matsuda index [10].
The current secondary analysis pertains to insulin clearance, assessed with fasting C-peptide-to-insulin ratio at the outset of each OGTT. Fasting C-peptide-to-insulin ratio is a validated measure of insulin clearance based on the assumptions that insulin and C-peptide are secreted in equimolar amounts and first-pass hepatic C-peptide clearance is negligible [1,2]. As shown in Fig. 1A, there were no differences in fasting C-peptide-to-insulin ratio between the liraglutide and placebo arms at randomization or at any of 12, 24, 36, or 48 weeks follow-up (all P non-significant). Similarly, after adjustment for its baseline measurement at randomization, fasting C-peptide-to-insulin ratio did not differ between the groups at any of these subsequent visits (all P non-significant) (Fig. 1B).
Comparison of liraglutide and placebo arms over 48-weeks with respect to (A) fasting C-peptide-to-insulin ratio, (B) baseline-adjusted fasting C-peptide-to-insulin ratio, (C) baseline-adjusted fasting C-peptide:insulin ratio further adjusted for change in weight up to indicated timepoint, and (D) baseline-adjusted fasting C-peptide:insulin ratio further adjusted for change in Matsuda index up to indicated timepoint. Panel (A) shows mean with standard deviation, with between-group comparisons made by analysis of variance. Panels (B), (C), and (D) show mean adjusted value with 95% confidence interval, with between-group comparisons made by analysis of covariance.
Furthermore, baseline-adjusted fasting C-peptide-to-insulin also did not differ between the groups after further adjustment for change in weight up to each visit (Fig. 1C) and change in Matsuda index up to each visit (all P non-significant) (Fig. 1D). In addition, the ratio of the area under the curve (AUC) for C-peptide to that of insulin (AUCC-peptide:AUCinsulin) did not differ between the liraglutide and placebo groups on the OGTTs at 12, 24, 36, or 48 weeks (data not shown).
Recent interest in the contribution of decreased insulin clearance to the pathophysiology of T2DM has raised the question of its potential modification by diabetes medications. Potential reasons for the conflicting findings on the impact of GLP-1 receptor agonists [3-9] have included physiologic differences between rodents and humans, observational study designs and modest sample sizes. Accordingly, this question would be best addressed in a clinical trial wherein patients with T2DM are randomly assigned to either GLP-1 receptor agonist or placebo.
However, to our knowledge, there have only been two such trials to date. In a crossover trial in which 30 patients with T2DM and coronary artery disease received 12-week courses of liraglutide and placebo in random sequence, insulin clearance did not differ between the interventions, though there was a nonsignificant trend towards higher clearance with liraglutide in completer analyses consisting of only 73% of 41 randomized participants [9]. In a parallel-arm trial in which 49 participants with prediabetes were randomized to 14 weeks treatment with either liraglutide or placebo [8], liraglutide decreased insulin clearance, although these analyses included only 72% of the 68 randomized participants. Thus, neither of these trials has conclusively defined the impact of liraglutide on insulin clearance.
In contrast with the aforementioned trials [8,9], all randomized participants were included in the current secondary analysis of the LIBRA trial and were treated for a much longer duration (48 weeks vs. 12–14 weeks), during which they underwent serial assessment of insulin clearance on four occasions. With these design strengths, the current study yielded very consistent findings, showing no impact of liraglutide on fasting C-peptide-to-insulin ratio on any of the four assessments over 48 weeks of treatment.
A limitation of this study is that fasting C-peptide-to-insulin ratio is a surrogate measure of total systemic insulin clearance consisting of both hepatic and peripheral extraction, the relative contributions of which cannot be ascertained with this index. Although findings were unaffected by adjustment for Matsuda index (which also has the statistical limitation of sharing a measurement [fasting insulin] with C-peptide-to-insulin ratio), definitive conclusions on the effect of liraglutide on hepatic insulin clearance cannot be drawn. Also, since the last dose of liraglutide was administered before breakfast on the day before each OGTT, the current findings should be interpreted as on-treatment effects. An additional limitation is that this trial was not specifically powered for this outcome measure, making these analyses exploratory by definition. The modest sample size raises the possibility of a type II error. Conversely, the sample size exceeds that of the earlier trials and was sufficient to demonstrate significant effects of liraglutide on β-cell function (by four different indices), glycemic control (A1c, 2-hour glucose on OGTT, and glucose tolerance), and BMI [10]. Moreover, our findings remained fully consistent on four assessments over 1 year. Taken together with the relative strengths of this analysis versus existing studies addressing this question, our findings suggest that chronic treatment with liraglutide may not impact insulin clearance early in the course of T2DM.
Notes
CONFLICTS OF INTEREST
Ravi Retnakaran holds the Boehringer Ingelheim Chair in Beta-cell Preservation, Function and Regeneration at Mount Sinai Hospital and his research program is supported by the Sun Life Financial Program to Prevent Diabetes in Women.
Ravi Retnakaran has received research funding and consulting honoraria from Novo Nordisk. Miya Boelling, Jiajie Pu, and Junwei Shen have nothing to disclose.