original article
Efficacy and safety of initial combination therapy with alogliptin plus metformin versus either as monotherapy in drug-naı¨ve patients with type 2 diabetes: a randomized, double-blind, 6-month study† R. E. Pratley1 , P. Fleck2 & C. Wilson2 1 Florida Hospital, Sanford Burnham Medical Research Institute, Orlando, FL, USA 2 Takeda Development Center Americas, Inc., Deerfield, IL, USA
Aim: To evaluate the efficacy and safety of the dipeptidyl peptidase-4 inhibitor alogliptin plus metformin (A + M) initial combination therapy versus either as monotherapy in drug-na¨ıve T2DM patients. Methods: This international, randomized, double-blind, placebo-controlled, 26-week study involved T2DM patients with hyperglycaemia (HbA1c 7.5–10.0%) following diet/exercise therapy. Patients (N = 784) received placebo, alogliptin (A, 12.5 mg BID or 25 mg QD), metformin (M, 500 or 1000 mg BID) or A + M (12.5/500 or 12.5/1000 mg BID); placebo, A25 for secondary analyses only. Endpoints: week 26 changes from baseline in HbA1c (primary), fasting plasma glucose (FPG) and 2-h postprandial glucose (PPG); incidences of clinical response and hyperglycaemic rescue. Results: Week 26 mean HbA1c reductions from baseline (8.45%) were −1.22 and −1.55% with A + M 12.5/500 and 12.5/1000 versus −0.56, −0.65, and −1.11% with A12.5, M500 and M1000 (p8.5%) and geographic region. Hyperglycaemic rescue was permitted for patients with a single fasting plasma glucose (FPG) level ≥225 to ≤275 mg/dl (≥12.49 to ≤15.27 mmol/l) depending on the timing of occurrence, or an HbA1c ≥ 8.5% and a ≤0.5% reduction from baseline (week 12 to the end of treatment). Rescue medication was a sulfonylurea (specific type/dose chosen by the investigator), which was added to the double-blind study medication regimen. Rescued patients continued with their assigned study medications and all other study-related activities. If a sulfonylurea was contraindicated or inappropriate, other rescue medications were prescribed at the investigator’s discretion; these patients discontinued study drug but continued with visits and procedures. The study also included prespecified pancreatitis, hepatic and renal safety withdrawal criteria. An independent data monitoring committee periodically reviewed the safety data, and an independent cardiovascular (CV) event adjudication committee conducted a prospective blinded review and adjudication of all deaths and all serious and selected nonserious potential CV events.
Ethics This study was designed in compliance with the Declaration of Helsinki, the International Conference on Harmonisation Harmonised Tripartite Guideline for Good Clinical Practice, and all applicable regional laws and regulations. All patients signed informed consent prior to participating in any study-related procedures. An institutional review board or ethics committee granted initial approval and performed ongoing reviews of this study throughout its duration.
Endpoints The primary efficacy endpoint was HbA1c change from baseline at week 26 (or at time of discontinuation of double-blind study medication or hyperglycaemic rescue). Changes in HbA1c and FPG over time were secondary endpoints, and clinical response (incidence of HbA1c < 7.0%), changes from baseline in 2-h postprandial glucose (PPG) and hyperglycaemic rescue were exploratory endpoints. Two-hour PPG tests were conducted only at selected sites capable of implementing them. At baseline/day 1, patients were not given study drug before or during PPG testing, whereas at the remaining visits, study drug was administered according to normal dosing routines at the start of the standardized prespecified 500 kcal liquid meal containing approximately 60% carbohydrate (or nutritionally equivalent/locally available liquid meal), which was consumed within 15 min.
Volume 16 No. 7 July 2014
original article
DIABETES, OBESITY AND METABOLISM
Other exploratory endpoints included weight; fasting measures of pancreatic β-cell function [proinsulin, insulin, proinsulin/insulin ratio and homeostasis model assessment beta-cell function (HOMA-BCF)]; lipids [total, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) cholesterol, and triglycerides]; and high-sensitivity C-reactive protein (hsCRP). Laboratory tests for the evaluation of these endpoints, as applicable, were conducted via blood draws following a minimum 8-h fast. Safety endpoints included incidence of adverse events (AEs), incidence of hypoglycaemic episodes, clinical laboratory tests (serum chemistry, haematology, urinalysis), vital sign measurements, physical examinations and 12-lead electrocardiograms. CV events, skin reactions and pancreatitis were examined as AEs of special interest.
Statistics The primary efficacy analysis involved statistical comparisons of the alogliptin plus metformin combination therapy regimens with their component monotherapy regimens – 12.5/500 BID versus alogliptin 12.5 BID and versus metformin 500 BID, and 12.5/1000 BID versus alogliptin 12.5 BID and versus metformin 1000 BID. For each set of comparisons, the null hypothesis of no additional effect was rejected only if both comparisons between a combination and its constituent doses were statistically significant at the two-sided 2.5% level, ensuring the overall false rejection level for the primary analysis was maintained at the 5% level. The primary analysis was conducted using the full analysis set (FAS) and an analysis of covariance model, with HbA1c change from baseline at week 26 (or at time of discontinuation of double-blind study medication or hyperglycaemic rescue) as the response variable, treatment and geographic region as fixed effects and baseline HbA1c as a continuous covariate. Each comparison was performed using contrasts derived from this model. Secondary efficacy analyses using the same methods as for the primary were also conducted to compare alogliptin 12.5 mg BID versus 25 mg QD and both combination regimens versus placebo. Data included in these analyses were collected on or after baseline and within 1 day (7 days for HbA1c) after the last dose of double-blind study medication unless a patient was rescued for hyperglycaemia, in which case only data collected on or prior to the date of rescue were used. At each visit, endpoints were analysed using the value collected at that visit or the last prior postbaseline value. Clinical response, hyperglycaemia and hyperglycaemic rescue incidences were analysed using the FAS and a logistic regression model, with effects for treatment, geographic region, study schedule and baseline HbA1c. A total of 105 patients per group ensured at least 90% power to declare whether either of the combination regimens was statistically superior to its constituent doses of alogliptin and metformin, assuming a 0.55% treatment effect between a combination and its constituent doses, a standard deviation of 1.0% and a two-sided false rejection rate of 2.5% for both comparisons between a combination and its constituents.
Volume 16 No. 7 July 2014
Results and Discussion Patient Population A total of 2478 patients were enrolled into the treatment period across 198 sites worldwide. Of these, 1694 patients failed screening or stabilization, whereas 784 were randomized to treatment. The majority of patients (60% overall) entered with a baseline HbA1c of 8.5% or lower. Most patients completed treatment, including 83% in both combination therapy groups, 83 and 86% in the metformin 500 and 1000 BID groups, 63 and 80% in the alogliptin 12.5 BID and 25 QD groups and 68% in the placebo group. Study drug discontinuation was mostly due to ‘voluntary withdrawal’ (8.4%), ‘adverse event’ (4.6%), and ‘lost to follow-up’ (3.8%), with the majority of voluntary withdrawals due to personal reasons (Table S1, Figure S2). Study drug compliance during the treatment period was high (overall mean of 96%). No noteworthy differences in demographics or baseline characteristics were observed among the treatment groups (Table 1). Concomitant medication use was largely consistent with the most common concurrent medical conditions, which included hypertension (50% of patients), dyslipidaemia (16%), postmenopause (15%), obesity (13%) and hyperlipidaemia (12%).
Glycaemic Control Efficacy Initial combination therapy with alogliptin plus metformin was significantly more efficacious versus alogliptin or metformin monotherapy in patients whose T2DM was inadequately controlled with diet and exercise alone. Differences in HbA1c between combination therapy and monotherapy were apparent from week 4 and continued throughout the study (Figure 1). At week 26, least squares (LS) mean [standard error (s.e.)] HbA1c changes from baseline were −1.22% (0.094) and −1.55% (0.090) with 12.5/500 and 12.5/1000 BID combination therapies, respectively, versus −0.56% (0.093) with alogliptin 12.5 BID, and −0.65% (0.094) and −1.11% (0.092) with metformin 500 and 1000 BID monotherapies (p < 0.001 for all comparisons of combination therapy vs. component monotherapies) (Figure 1). Subgroup analyses demonstrated clinically meaningful HbA1c reductions regardless of baseline HbA1c, age, gender, race, ethnicity and baseline BMI. Significantly more patients also achieved the target goal of HbA1c < 7.0% by week 26: 47.1 and 59.5% with 12.5/500 and 12.5/1000 BID versus 20.2% with alogliptin 12.5 BID and 27.2 and 34.3% with metformin 500 and 1000 BID, respectively (p < 0.01 for all comparisons of combination therapy vs. component monotherapies) (Figure S3). Changes in FPG followed the same pattern that was observed for HbA1c. Combination versus monotherapy treatment differences were apparent from week 1 and continued throughout the study (p < 0.05 for all comparisons at all time points, except 12.5/1000 BID vs. metformin 1000 BID at week 20). At week 26, LS mean (s.e.) FPG reductions from baseline were −1.76 (0.25) and −2.55 (0.24) mg/dl with 12.5/500 and 12.5/1000 BID, respectively, compared
doi:10.1111/dom.12258 615
616 Pratley et al. 52.6 (9.38) 103 (92.0) 9 (8.0) 48 (42.9) 64 (57.1) 17 (15.2) 3 (2.7) 84 (75.0) 8 (7.1) 30.8 (5.22) 30.5 (20.0–44.8) 3.6 (4.12) 2.4 (0.2–20.4)
100 (91.7) 9 (8.3)
55 (50.5) 54 (49.5)
20 (18.3) 8 (7.3)
76 (69.7) 5 (4.6)
31.2 (5.27) 30.4 (20.8–43.7)
4.3 (4.78) 2.7 (0.2–31.4)
4.0 (4.80) 2.2 (0.0–28.6)
30.4 (5.16) 29.4 (20.8–45.9)
83 (73.5) 6 (5.3)
21 (18.6) 3 (2.7)
63 (55.8) 50 (44.2)
96 (85.0) 17 (15.0)
53.7 (9.70)
3.8 (3.90) 2.4 (0.0–18.3)
30.2 (4.84) 29.0 (21.1–43.9)
85 (74.6) 4 (3.5)
19 (16.7) 6 (5.3)
47 (41.2) 67 (58.8)
95 (83.3) 19 (16.7)
54.6 (10.20)
4.1 (4.59) 2.8 (0.1–27.1)
30.5 (5.0) 29.6 (20.7–43.3)
79 (71.2) 6 (5.4)
20 (18.0) 6 (5.4)
51 (45.9) 60 (54.1)
94 (84.7) 17 (15.3)
52.6 (11.30)
1000 BID (n = 111)
500 BID (n = 114)
12.5 BID (n = 113)
25 QD (n = 112)
53.1 (9.60)
Placebo (n = 109)
BMI, body mass index; s.d., standard deviation. *Includes American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander and Multiracial.
Age Mean (s.d.), years
Efficacy and safety of initial combination therapy with alogliptin plus metformin versus either as monotherapy in drug-naı¨ve patients with type 2 diabetes: a randomized, double-blind, 6-month study† R. E. Pratley1 , P. Fleck2 & C. Wilson2 1 Florida Hospital, Sanford Burnham Medical Research Institute, Orlando, FL, USA 2 Takeda Development Center Americas, Inc., Deerfield, IL, USA
Aim: To evaluate the efficacy and safety of the dipeptidyl peptidase-4 inhibitor alogliptin plus metformin (A + M) initial combination therapy versus either as monotherapy in drug-na¨ıve T2DM patients. Methods: This international, randomized, double-blind, placebo-controlled, 26-week study involved T2DM patients with hyperglycaemia (HbA1c 7.5–10.0%) following diet/exercise therapy. Patients (N = 784) received placebo, alogliptin (A, 12.5 mg BID or 25 mg QD), metformin (M, 500 or 1000 mg BID) or A + M (12.5/500 or 12.5/1000 mg BID); placebo, A25 for secondary analyses only. Endpoints: week 26 changes from baseline in HbA1c (primary), fasting plasma glucose (FPG) and 2-h postprandial glucose (PPG); incidences of clinical response and hyperglycaemic rescue. Results: Week 26 mean HbA1c reductions from baseline (8.45%) were −1.22 and −1.55% with A + M 12.5/500 and 12.5/1000 versus −0.56, −0.65, and −1.11% with A12.5, M500 and M1000 (p8.5%) and geographic region. Hyperglycaemic rescue was permitted for patients with a single fasting plasma glucose (FPG) level ≥225 to ≤275 mg/dl (≥12.49 to ≤15.27 mmol/l) depending on the timing of occurrence, or an HbA1c ≥ 8.5% and a ≤0.5% reduction from baseline (week 12 to the end of treatment). Rescue medication was a sulfonylurea (specific type/dose chosen by the investigator), which was added to the double-blind study medication regimen. Rescued patients continued with their assigned study medications and all other study-related activities. If a sulfonylurea was contraindicated or inappropriate, other rescue medications were prescribed at the investigator’s discretion; these patients discontinued study drug but continued with visits and procedures. The study also included prespecified pancreatitis, hepatic and renal safety withdrawal criteria. An independent data monitoring committee periodically reviewed the safety data, and an independent cardiovascular (CV) event adjudication committee conducted a prospective blinded review and adjudication of all deaths and all serious and selected nonserious potential CV events.
Ethics This study was designed in compliance with the Declaration of Helsinki, the International Conference on Harmonisation Harmonised Tripartite Guideline for Good Clinical Practice, and all applicable regional laws and regulations. All patients signed informed consent prior to participating in any study-related procedures. An institutional review board or ethics committee granted initial approval and performed ongoing reviews of this study throughout its duration.
Endpoints The primary efficacy endpoint was HbA1c change from baseline at week 26 (or at time of discontinuation of double-blind study medication or hyperglycaemic rescue). Changes in HbA1c and FPG over time were secondary endpoints, and clinical response (incidence of HbA1c < 7.0%), changes from baseline in 2-h postprandial glucose (PPG) and hyperglycaemic rescue were exploratory endpoints. Two-hour PPG tests were conducted only at selected sites capable of implementing them. At baseline/day 1, patients were not given study drug before or during PPG testing, whereas at the remaining visits, study drug was administered according to normal dosing routines at the start of the standardized prespecified 500 kcal liquid meal containing approximately 60% carbohydrate (or nutritionally equivalent/locally available liquid meal), which was consumed within 15 min.
Volume 16 No. 7 July 2014
original article
DIABETES, OBESITY AND METABOLISM
Other exploratory endpoints included weight; fasting measures of pancreatic β-cell function [proinsulin, insulin, proinsulin/insulin ratio and homeostasis model assessment beta-cell function (HOMA-BCF)]; lipids [total, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) cholesterol, and triglycerides]; and high-sensitivity C-reactive protein (hsCRP). Laboratory tests for the evaluation of these endpoints, as applicable, were conducted via blood draws following a minimum 8-h fast. Safety endpoints included incidence of adverse events (AEs), incidence of hypoglycaemic episodes, clinical laboratory tests (serum chemistry, haematology, urinalysis), vital sign measurements, physical examinations and 12-lead electrocardiograms. CV events, skin reactions and pancreatitis were examined as AEs of special interest.
Statistics The primary efficacy analysis involved statistical comparisons of the alogliptin plus metformin combination therapy regimens with their component monotherapy regimens – 12.5/500 BID versus alogliptin 12.5 BID and versus metformin 500 BID, and 12.5/1000 BID versus alogliptin 12.5 BID and versus metformin 1000 BID. For each set of comparisons, the null hypothesis of no additional effect was rejected only if both comparisons between a combination and its constituent doses were statistically significant at the two-sided 2.5% level, ensuring the overall false rejection level for the primary analysis was maintained at the 5% level. The primary analysis was conducted using the full analysis set (FAS) and an analysis of covariance model, with HbA1c change from baseline at week 26 (or at time of discontinuation of double-blind study medication or hyperglycaemic rescue) as the response variable, treatment and geographic region as fixed effects and baseline HbA1c as a continuous covariate. Each comparison was performed using contrasts derived from this model. Secondary efficacy analyses using the same methods as for the primary were also conducted to compare alogliptin 12.5 mg BID versus 25 mg QD and both combination regimens versus placebo. Data included in these analyses were collected on or after baseline and within 1 day (7 days for HbA1c) after the last dose of double-blind study medication unless a patient was rescued for hyperglycaemia, in which case only data collected on or prior to the date of rescue were used. At each visit, endpoints were analysed using the value collected at that visit or the last prior postbaseline value. Clinical response, hyperglycaemia and hyperglycaemic rescue incidences were analysed using the FAS and a logistic regression model, with effects for treatment, geographic region, study schedule and baseline HbA1c. A total of 105 patients per group ensured at least 90% power to declare whether either of the combination regimens was statistically superior to its constituent doses of alogliptin and metformin, assuming a 0.55% treatment effect between a combination and its constituent doses, a standard deviation of 1.0% and a two-sided false rejection rate of 2.5% for both comparisons between a combination and its constituents.
Volume 16 No. 7 July 2014
Results and Discussion Patient Population A total of 2478 patients were enrolled into the treatment period across 198 sites worldwide. Of these, 1694 patients failed screening or stabilization, whereas 784 were randomized to treatment. The majority of patients (60% overall) entered with a baseline HbA1c of 8.5% or lower. Most patients completed treatment, including 83% in both combination therapy groups, 83 and 86% in the metformin 500 and 1000 BID groups, 63 and 80% in the alogliptin 12.5 BID and 25 QD groups and 68% in the placebo group. Study drug discontinuation was mostly due to ‘voluntary withdrawal’ (8.4%), ‘adverse event’ (4.6%), and ‘lost to follow-up’ (3.8%), with the majority of voluntary withdrawals due to personal reasons (Table S1, Figure S2). Study drug compliance during the treatment period was high (overall mean of 96%). No noteworthy differences in demographics or baseline characteristics were observed among the treatment groups (Table 1). Concomitant medication use was largely consistent with the most common concurrent medical conditions, which included hypertension (50% of patients), dyslipidaemia (16%), postmenopause (15%), obesity (13%) and hyperlipidaemia (12%).
Glycaemic Control Efficacy Initial combination therapy with alogliptin plus metformin was significantly more efficacious versus alogliptin or metformin monotherapy in patients whose T2DM was inadequately controlled with diet and exercise alone. Differences in HbA1c between combination therapy and monotherapy were apparent from week 4 and continued throughout the study (Figure 1). At week 26, least squares (LS) mean [standard error (s.e.)] HbA1c changes from baseline were −1.22% (0.094) and −1.55% (0.090) with 12.5/500 and 12.5/1000 BID combination therapies, respectively, versus −0.56% (0.093) with alogliptin 12.5 BID, and −0.65% (0.094) and −1.11% (0.092) with metformin 500 and 1000 BID monotherapies (p < 0.001 for all comparisons of combination therapy vs. component monotherapies) (Figure 1). Subgroup analyses demonstrated clinically meaningful HbA1c reductions regardless of baseline HbA1c, age, gender, race, ethnicity and baseline BMI. Significantly more patients also achieved the target goal of HbA1c < 7.0% by week 26: 47.1 and 59.5% with 12.5/500 and 12.5/1000 BID versus 20.2% with alogliptin 12.5 BID and 27.2 and 34.3% with metformin 500 and 1000 BID, respectively (p < 0.01 for all comparisons of combination therapy vs. component monotherapies) (Figure S3). Changes in FPG followed the same pattern that was observed for HbA1c. Combination versus monotherapy treatment differences were apparent from week 1 and continued throughout the study (p < 0.05 for all comparisons at all time points, except 12.5/1000 BID vs. metformin 1000 BID at week 20). At week 26, LS mean (s.e.) FPG reductions from baseline were −1.76 (0.25) and −2.55 (0.24) mg/dl with 12.5/500 and 12.5/1000 BID, respectively, compared
doi:10.1111/dom.12258 615
616 Pratley et al. 52.6 (9.38) 103 (92.0) 9 (8.0) 48 (42.9) 64 (57.1) 17 (15.2) 3 (2.7) 84 (75.0) 8 (7.1) 30.8 (5.22) 30.5 (20.0–44.8) 3.6 (4.12) 2.4 (0.2–20.4)
100 (91.7) 9 (8.3)
55 (50.5) 54 (49.5)
20 (18.3) 8 (7.3)
76 (69.7) 5 (4.6)
31.2 (5.27) 30.4 (20.8–43.7)
4.3 (4.78) 2.7 (0.2–31.4)
4.0 (4.80) 2.2 (0.0–28.6)
30.4 (5.16) 29.4 (20.8–45.9)
83 (73.5) 6 (5.3)
21 (18.6) 3 (2.7)
63 (55.8) 50 (44.2)
96 (85.0) 17 (15.0)
53.7 (9.70)
3.8 (3.90) 2.4 (0.0–18.3)
30.2 (4.84) 29.0 (21.1–43.9)
85 (74.6) 4 (3.5)
19 (16.7) 6 (5.3)
47 (41.2) 67 (58.8)
95 (83.3) 19 (16.7)
54.6 (10.20)
4.1 (4.59) 2.8 (0.1–27.1)
30.5 (5.0) 29.6 (20.7–43.3)
79 (71.2) 6 (5.4)
20 (18.0) 6 (5.4)
51 (45.9) 60 (54.1)
94 (84.7) 17 (15.3)
52.6 (11.30)
1000 BID (n = 111)
500 BID (n = 114)
12.5 BID (n = 113)
25 QD (n = 112)
53.1 (9.60)
Placebo (n = 109)
BMI, body mass index; s.d., standard deviation. *Includes American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander and Multiracial.
Age Mean (s.d.), years
