research letter
Diabetes, Obesity and Metabolism 16: 1036–1039, 2014. © 2014 John Wiley & Sons Ltd
research letter
The dipeptidyl peptidase-4 inhibitor linagliptin lowers postprandial glucose and improves measures of β-cell function in type 2 diabetes†
Progressive deterioration of pancreatic β-cell function in patients with type 2 diabetes mellitus (T2DM) contributes to worsening of hyperglycaemia. To investigate the effects of the dipeptidyl peptidase-4 inhibitor linagliptin on β-cell function parameters, a pooled analysis of six randomized, 24-week, placebo-controlled, phase 3 trials of 5 mg of linagliptin daily was performed in 2701 patients with T2DM (linagliptin, n = 1905; placebo, n = 796). At week 24, observed improvements in HbA1c, fasting plasma glucose, and 2-h postprandial glucose were significantly greater for linagliptin than placebo (all p < 0.0001). Homeostasis model assessment (HOMA)-%β, as a surrogate marker of fasting β-cell function, was significantly improved with linagliptin, and did not change with placebo (placebo-adjusted mean ± s.e. change for linagliptin: 16.5 ± 4.6 (mU/l)/(mmol/l); p = 0.0003). Further study is required to determine if the significant improvement in HOMA-%β with linagliptin will translate into long-term improvements in β-cell function. Keywords: beta cell, DPP-IV inhibitor, drug mechanism, glycaemic control, type 2 diabetes Date submitted 23 December 2013; date of first decision 6 February 2014; date of final acceptance 2 May 2014
Introduction Progressive deterioration of pancreatic β-cell function in patients with type 2 diabetes mellitus (T2DM) contributes to glycaemia deterioration. In the UK Prospective Diabetes Study [1] and A Diabetes Outcome Progression Trial [2], increasing hyperglycaemia over time despite therapy with a sulphonylurea and/or metformin was associated with decreasing β-cell function. The incretins glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) are key gastrointestinal hormones that regulate pancreatic α- and βcell functions [3]. Treatment with incretin-based therapies may therefore improve β-cell function and potentially delay or reverse disease progression [4]. Linagliptin is a once-daily, orally available, selective inhibitor of dipeptidyl peptidase (DPP)-4, the enzyme responsible for cleavage of GLP-1 and GIP [5]. Following administration, linagliptin increases the availability of active GLP-1 and GIP which, in turn, stimulate glucose-dependent insulin release from pancreatic β-cells (GLP-1 and GIP) while reducing glucagon output from pancreatic α-cells (GLP-1 only) [5]. In phase 3 clinical studies, linagliptin either as monotherapy or in combination with other oral antidiabetes drugs produced clinically significant improvements in glycated haemoglobin Correspondence to: Dr Tim Heise, Profil Institut fur ¨ Stoffwechselforschung GmbH, Hellersbergstr. 9, Neuss, Germany. E-mail: tim.heise@profil.com † A poster based on this study was presented at the 73rd Scientific Sessions of the American Diabetes Association, 21–25 June 2013, Chicago, IL, USA.
(HbA1c), fasting plasma glucose (FPG) and postprandial glucose (PPG). In preclinical studies, linagliptin improved β-cell function via GLP-1 stabilization as well as delayed the onset of diabetes by inhibiting the destruction of β-cells [6,7]. The purpose of this analysis was to evaluate the effect of linagliptin on surrogate measures of β-cell function using pooled data from six phase 3 trials of T2DM patients on various glucose-lowering background therapies.
Methods This pooled analysis included six randomized, 24-week, placebo-controlled, phase 3 trials of T2DM patients who received 5 mg of linagliptin as monotherapy (NCT00621140; NCT01194830), as add-on to metformin (NCT00601250), metformin plus sulphonylurea (NCT00602472) or metformin plus pioglitazone (NCT00996658), and as initial combination therapy with metformin (NCT00798161). These studies were selected because the study participants had valid homeostasis model assessment (HOMA)-%β measurements. In addition, the close similarity of the study designs enabled pooling of the respective data to perform an integrated comparison of the treatment regimens on surrogate measures of β-cell function. The appropriateness of pooling the data was examined statistically using a generalized linear model (for adjusted differences) with the following factors: study, baseline HOMA%β, baseline HbA1c, treatment, and interaction of treatment by study. No significant interaction effects were detected.
research letter
DIABETES, OBESITY AND METABOLISM
All data analyses were performed using descriptive statistics [mean ± standard deviation (s.d.) for baseline parameters, and mean ± standard error (s.e.) for on-treatment comparisons] ® ® compiled by sas version 9.2 or higher (SAS Institute, Cary, NC, USA). All patients who received one or more doses of the study drug (treated set) and who had a valid HOMA-%β measurement were included for analysis. HOMA%β, a surrogate marker for β-cell function, is modelled on simultaneous measures of fasting glucose and insulin levels [8]. Analysis of covariance (ancova) models were used to analyse changes in HOMA-%β index and in 2-h PPG levels after a meal tolerance test, as well as HbA1c and FPG level data. The statistical model included treatment and study as fixed classification effects, and baseline HbA1c as linear covariate. For non-HbA1c analyses, the respective baseline value (e.g. 2-h PPG) was added as the covariate. The analysis was performed on the full analysis set (FAS), defined as all randomized patients who received one or more doses of the study drug and had a baseline HbA1c measurement and one or more on-treatment HbA1c measurements. For HbA1c and FPG, on-treatment values were assessed with missing data imputed using a last observation carried forward (LOCF) approach. Because HOMA-%β and PPG were measured at baseline and week 24, an observed cases (OC) approach without imputation of missing data was used for HOMA%β and PPG analysis. Meal tolerance tests for PPG were measured in a subset of patients who participated in four of
the six trials (NCT00621140; NCT01194830; NCT00601250; NCT00798161).
Results Data from 2701 patients were analysed (treated set: linagliptin, n = 1905; placebo, n = 796). Baseline demographics were similar between the treatment groups. There was an equal proportion of men and women, and mean age was 57 ± 10 years. Half of the patients were white, 41% were Asian and 9% were black/African American. No differences between the linagliptin and placebo groups were observed at baseline, respectively, with respect to body mass index (29 ± 5 kg/m2 vs. 30 ± 5 kg/m2 ), HbA1c (both 8.2 ± 0.9%), HOMA-%β [54.6 ± 101.6 (mU/l)/(mmol/l) vs. 49.4 ± 47.7 (mU/l)/(mmol/l)], FPG (9.1 ± 2.4 mmol/l vs. 9.3 ± 2.5 mmol/l), PPG (14.8 ± 4.1 mmol/l vs. 14.1 ± 4.1 mmol/l), and proportion of patients with T2DM of >5 years duration (56 vs. 54%). A total of 238 patients prematurely discontinued study medication before week 24, including 149 (7.8%) in the linagliptin group and 89 (11.2%) in the placebo group. Exposure to linagliptin and placebo was 165 ± 30 and 162 ± 34 days, respectively. At week 24, the observed improvements in HbA1c, FPG and PPG were significantly greater with linagliptin than with placebo (all p < 0.0001; Table 1). At week 24, HOMA-%β did not significantly change with placebo
Table 1. Change in parameters from baseline to week 24 (FAS). Parameter HbA1c (%) Patients, n Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.* Difference versus placebo, adjusted mean ± s.e. FPG (mmol/l) Patients, n† Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.* Difference versus placebo, adjusted mean ± s.e. PPG (mmol/l) Patients, n‡ Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.§ Difference versus placebo, adjusted mean ± s.e. HOMA-%β (mU/l)/(mmol/l) Patients, n¶ Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.§ Difference versus placebo, adjusted mean ± s.e.
Linagliptin
Placebo
p-Value
1865 8.17 ± 0.86 −0.63 ± 0.03 −0.63 ± 0.04
777 8.20 ± 0.92 0.01 ± 0.04 —
Diabetes, Obesity and Metabolism 16: 1036–1039, 2014. © 2014 John Wiley & Sons Ltd
research letter
The dipeptidyl peptidase-4 inhibitor linagliptin lowers postprandial glucose and improves measures of β-cell function in type 2 diabetes†
Progressive deterioration of pancreatic β-cell function in patients with type 2 diabetes mellitus (T2DM) contributes to worsening of hyperglycaemia. To investigate the effects of the dipeptidyl peptidase-4 inhibitor linagliptin on β-cell function parameters, a pooled analysis of six randomized, 24-week, placebo-controlled, phase 3 trials of 5 mg of linagliptin daily was performed in 2701 patients with T2DM (linagliptin, n = 1905; placebo, n = 796). At week 24, observed improvements in HbA1c, fasting plasma glucose, and 2-h postprandial glucose were significantly greater for linagliptin than placebo (all p < 0.0001). Homeostasis model assessment (HOMA)-%β, as a surrogate marker of fasting β-cell function, was significantly improved with linagliptin, and did not change with placebo (placebo-adjusted mean ± s.e. change for linagliptin: 16.5 ± 4.6 (mU/l)/(mmol/l); p = 0.0003). Further study is required to determine if the significant improvement in HOMA-%β with linagliptin will translate into long-term improvements in β-cell function. Keywords: beta cell, DPP-IV inhibitor, drug mechanism, glycaemic control, type 2 diabetes Date submitted 23 December 2013; date of first decision 6 February 2014; date of final acceptance 2 May 2014
Introduction Progressive deterioration of pancreatic β-cell function in patients with type 2 diabetes mellitus (T2DM) contributes to glycaemia deterioration. In the UK Prospective Diabetes Study [1] and A Diabetes Outcome Progression Trial [2], increasing hyperglycaemia over time despite therapy with a sulphonylurea and/or metformin was associated with decreasing β-cell function. The incretins glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) are key gastrointestinal hormones that regulate pancreatic α- and βcell functions [3]. Treatment with incretin-based therapies may therefore improve β-cell function and potentially delay or reverse disease progression [4]. Linagliptin is a once-daily, orally available, selective inhibitor of dipeptidyl peptidase (DPP)-4, the enzyme responsible for cleavage of GLP-1 and GIP [5]. Following administration, linagliptin increases the availability of active GLP-1 and GIP which, in turn, stimulate glucose-dependent insulin release from pancreatic β-cells (GLP-1 and GIP) while reducing glucagon output from pancreatic α-cells (GLP-1 only) [5]. In phase 3 clinical studies, linagliptin either as monotherapy or in combination with other oral antidiabetes drugs produced clinically significant improvements in glycated haemoglobin Correspondence to: Dr Tim Heise, Profil Institut fur ¨ Stoffwechselforschung GmbH, Hellersbergstr. 9, Neuss, Germany. E-mail: tim.heise@profil.com † A poster based on this study was presented at the 73rd Scientific Sessions of the American Diabetes Association, 21–25 June 2013, Chicago, IL, USA.
(HbA1c), fasting plasma glucose (FPG) and postprandial glucose (PPG). In preclinical studies, linagliptin improved β-cell function via GLP-1 stabilization as well as delayed the onset of diabetes by inhibiting the destruction of β-cells [6,7]. The purpose of this analysis was to evaluate the effect of linagliptin on surrogate measures of β-cell function using pooled data from six phase 3 trials of T2DM patients on various glucose-lowering background therapies.
Methods This pooled analysis included six randomized, 24-week, placebo-controlled, phase 3 trials of T2DM patients who received 5 mg of linagliptin as monotherapy (NCT00621140; NCT01194830), as add-on to metformin (NCT00601250), metformin plus sulphonylurea (NCT00602472) or metformin plus pioglitazone (NCT00996658), and as initial combination therapy with metformin (NCT00798161). These studies were selected because the study participants had valid homeostasis model assessment (HOMA)-%β measurements. In addition, the close similarity of the study designs enabled pooling of the respective data to perform an integrated comparison of the treatment regimens on surrogate measures of β-cell function. The appropriateness of pooling the data was examined statistically using a generalized linear model (for adjusted differences) with the following factors: study, baseline HOMA%β, baseline HbA1c, treatment, and interaction of treatment by study. No significant interaction effects were detected.
research letter
DIABETES, OBESITY AND METABOLISM
All data analyses were performed using descriptive statistics [mean ± standard deviation (s.d.) for baseline parameters, and mean ± standard error (s.e.) for on-treatment comparisons] ® ® compiled by sas version 9.2 or higher (SAS Institute, Cary, NC, USA). All patients who received one or more doses of the study drug (treated set) and who had a valid HOMA-%β measurement were included for analysis. HOMA%β, a surrogate marker for β-cell function, is modelled on simultaneous measures of fasting glucose and insulin levels [8]. Analysis of covariance (ancova) models were used to analyse changes in HOMA-%β index and in 2-h PPG levels after a meal tolerance test, as well as HbA1c and FPG level data. The statistical model included treatment and study as fixed classification effects, and baseline HbA1c as linear covariate. For non-HbA1c analyses, the respective baseline value (e.g. 2-h PPG) was added as the covariate. The analysis was performed on the full analysis set (FAS), defined as all randomized patients who received one or more doses of the study drug and had a baseline HbA1c measurement and one or more on-treatment HbA1c measurements. For HbA1c and FPG, on-treatment values were assessed with missing data imputed using a last observation carried forward (LOCF) approach. Because HOMA-%β and PPG were measured at baseline and week 24, an observed cases (OC) approach without imputation of missing data was used for HOMA%β and PPG analysis. Meal tolerance tests for PPG were measured in a subset of patients who participated in four of
the six trials (NCT00621140; NCT01194830; NCT00601250; NCT00798161).
Results Data from 2701 patients were analysed (treated set: linagliptin, n = 1905; placebo, n = 796). Baseline demographics were similar between the treatment groups. There was an equal proportion of men and women, and mean age was 57 ± 10 years. Half of the patients were white, 41% were Asian and 9% were black/African American. No differences between the linagliptin and placebo groups were observed at baseline, respectively, with respect to body mass index (29 ± 5 kg/m2 vs. 30 ± 5 kg/m2 ), HbA1c (both 8.2 ± 0.9%), HOMA-%β [54.6 ± 101.6 (mU/l)/(mmol/l) vs. 49.4 ± 47.7 (mU/l)/(mmol/l)], FPG (9.1 ± 2.4 mmol/l vs. 9.3 ± 2.5 mmol/l), PPG (14.8 ± 4.1 mmol/l vs. 14.1 ± 4.1 mmol/l), and proportion of patients with T2DM of >5 years duration (56 vs. 54%). A total of 238 patients prematurely discontinued study medication before week 24, including 149 (7.8%) in the linagliptin group and 89 (11.2%) in the placebo group. Exposure to linagliptin and placebo was 165 ± 30 and 162 ± 34 days, respectively. At week 24, the observed improvements in HbA1c, FPG and PPG were significantly greater with linagliptin than with placebo (all p < 0.0001; Table 1). At week 24, HOMA-%β did not significantly change with placebo
Table 1. Change in parameters from baseline to week 24 (FAS). Parameter HbA1c (%) Patients, n Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.* Difference versus placebo, adjusted mean ± s.e. FPG (mmol/l) Patients, n† Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.* Difference versus placebo, adjusted mean ± s.e. PPG (mmol/l) Patients, n‡ Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.§ Difference versus placebo, adjusted mean ± s.e. HOMA-%β (mU/l)/(mmol/l) Patients, n¶ Baseline, mean ± s.d. Change from baseline, adjusted mean ± s.e.§ Difference versus placebo, adjusted mean ± s.e.
Linagliptin
Placebo
p-Value
1865 8.17 ± 0.86 −0.63 ± 0.03 −0.63 ± 0.04
777 8.20 ± 0.92 0.01 ± 0.04 —
