T. Forst, B. Uhlig-Laske*, A. Ring*, U. Graefe-Mody*, C. Friedrich*, K. Herbach*, H.-J. Woerle* and K. A. Dugi

Transcription

1 Original Article: Treatment Linagliptin (BI 1356), a potent and selective DPP-4 inhibitor, is safe and efficacious in combination with metformin in patients with inadequately controlled Type 2 diabetes T. Forst, B. Uhlig-Laske*, A. Ring*, U. Graefe-Mody*, C. Friedrich*, K. Herbach*, H.-J. Woerle* and K. A. Dugi Institut für Klinische Forschung und Entwicklung GmbH, Mainz, *Boehringer Ingelheim Pharma GmbH and Co. KG, Biberach and Boehringer Ingelheim GmbH, Ingelheim, Germany Accepted 6 September 2010 DIABETICMedicine DOI: /j x Abstract Aims The efficacy and safety of the dipeptidyl peptidase-4 inhibitor, linagliptin, added to ongoing metformin therapy, were assessed in patients with Type 2 diabetes who had inadequate glycaemic control (HbA 1c 7.5 to 10%; 58.5 to 85.8 mmol mol) with metformin alone. Methods Patients (n = 333) were randomized to receive double-blind linagliptin (1, 5 or 10 mg once daily) or placebo or openlabel glimepiride (1 3 mg once daily). The primary outcome measure was the change from baseline in HbA 1c at week 12 in patients receiving combination therapy compared with metformin alone. Results Twelve weeks of treatment resulted in a mean (sem) placebo-corrected lowering in HbA 1c levels of 0.40% ( 0.14); 4.4 mmol mol ( 1.5) for 1 mg linagliptin, 0.73% ( 0.14); 8.0 mmol mol ( 1.5) for 5 mg, and 0.67% ( 0.14); 7.3 mmol mol ( 1.5) for 10 mg. Differences between linagliptin and placebo were statistically significant for all doses (1 mg, P = 0.01; 5 mg and 10 mg, P < ). The change in mean (sem) placebo-corrected HbA 1c from baseline was )0.90% ( 0.13); )9.8 mmol mol ( 1.4) for glimepiride. Adjusted and placebo-corrected mean changes in fasting plasma glucose were )1.1 mmol l for linagliptin 1 mg (P =0.002),)1.9 mmol l for 5 mg and )1.6 mmol l for 10 mg (both P < ). One hundred and six (43.1%) patients reported adverse events; the incidence was similar across all five groups. There were no hypoglycaemic events for linagliptin or placebo, whereas three patients (5%) receiving glimepiride experienced hypoglycaemia. Conclusions The addition of linagliptin to ongoing metformin treatment in patients with Type 2 diabetes was well tolerated and resulted in significant and clinically relevant improvements in glycaemic control, with 5 mg linagliptin being the most effective dose. Diabet. Med. 27, (2010) Keywords linagliptin, metformin, safety and efficacy Abbreviations DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide-1; HOMA, homeostasis model assessment; MedDRA, Medical Dictionary for Regulatory Activities Introduction Metformin is the most commonly used first-line treatment for Type 2 diabetes mellitus [1,2]. However, such treatment with monotherapy is often insufficient to achieve glycaemic control Correspondence to: Klaus A Dugi MD, Corporate Department of Medical Affairs, Boehringer Ingelheim GmbH, Binger Str. 173, D Ingelheim, Germany. klaus.dugi@boehringer-ingelheim.com [3]. Thus, combination therapy of several anti-hyperglycaemic agents is frequently required [2]. Effective treatment can be further complicated by the progressive nature of diabetes and its associated deterioration of glycaemic control over time, making it necessary to further augment treatment with additional therapies [1 4]. The dipeptidyl peptidase-4 (DPP-4) inhibitors are one of the newest options for treatment of Type 2 diabetes. These agents act by inhibiting the degradation of glucagon-like peptide-1 (GLP-1) Diabetic Medicine ª 2010 Diabetes UK 1409

2 DIABETICMedicine Linagliptin added to metformin in Type 2 diabetes T. Forst et al. and glucose-dependent insulinotropic peptide (GIP) by the plasma DPP-4 enzyme [5 7]. The consequent increase in exposure to these hormones serves to augment glucosedependent insulin release [8,9]. The glucagon response following a meal is also suppressed by GLP-1 [10]. DPP-4 inhibitors are currently considered as possible second- and thirdline agents for combination with metformin that provide improved glycaemic control over metformin alone. The effect of DPP-4 inhibition on exposure to GLP-1, i.e. the lowering of circulating glucose through enhanced insulin secretion and inhibition of glucagon secretion, is believed to complement the suppression of hepatic glucose production and improved insulin sensitivity associated with metformin [11,12]. Linagliptin (BI 1356) is an orally administered, potent, selective xanthine-based DPP-4 inhibitor for Type 2 diabetes [13 15]. Linagliptin shows similar maximal efficacy for inhibition of DPP-4 in vitro when compared with other DPP-4 inhibitors, but with higher potency (IC 50 = 1 nm for linagliptin vs. 19, 62, 50 and 24 nm, for sitagliptin, vildagliptin, saxagliptin and alogliptin, respectively) [16]. It has excellent selectivity for DPP-4 vs. DPP-8 (40,000-fold) and DPP-9 (> 10,000-fold). It shows little interaction with other proteases (prolyloligopeptidase, aminopeptidase N, aminopeptidase P, trypsin, plasmin, thrombin; > 10,000-fold selectivity) [16], has low affinity for herg channel (88% current remaining at 1 lm), and has no significant inhibitory effect on the cytochrome P450 (CYP 450) enzymes (IC 50 >50lm) [17 19]. The long duration of action of linagliptin makes it suitable for once-daily dosing without dose titration [16,20]. A pharmacokinetic study in 47 male patients with Type 2 diabetes found that linagliptin shows a short accumulation half-life ( h), rapidly attained steady state (2 5 days) with little accumulation (range: ). The long terminal half-life ( h) produced a sustained inhibition of DPP-4, which correlated with plasma linagliptin concentrations [15]. In a multiple-dose study in healthy subjects, co-administration of linagliptin with metformin did not have a clinically relevant effect on the pharmacokinetics or pharmacodynamics of either agent. This suggests that linagliptin and metformin can safely be administered concomitantly without dose adjustment [21]. The purpose of the present study was to test the hypothesis that linagliptin is safe and efficacious in patients with Type 2 diabetes failing to achieve glycaemic control despite therapy with metformin. Patients and methods Study population Male and female patients diagnosed with Type 2 diabetes for at least 3 months, aged 21 to 75 years, with a BMI of 25 to 40 kg m 2, were eligible for this study. At the time of screening, all patients had inadequate glycaemic control despite having been previously treated with metformin alone or with metformin and one other oral hypoglycaemic agent other than rosiglitazone or pioglitazone (the anti-diabetic therapy had to be unchanged for 10 weeks prior to screening). For patients previously treated with metformin and one other oral anti-diabetic drug, inadequate glycaemic control was defined as an HbA 1c level from % ( mmol mol) at screening; for patients previously treated with metformin alone, inadequate glycaemic control was defined as an HbA 1c level from % ( mmol mol) at screening. Patients were excluded if they had fasting plasma glucose concentrations of > 13.3 mmol l (measured on two separate days) or if they had been treated with rosiglitazone or pioglitazone within 6 months, or with insulin in the 3 months, prior to screening. Patients were also ineligible if they had experienced a clinically relevant cardiovascular disease, myocardial infarction, stroke or transient ischaemic attack within 6 months before enrolment, or if they had one or more of a list of specified clinical laboratory abnormalities. Study design This was a 12-week, multi-centre, randomized, double-blind, placebo-controlled, five parallel group study that compared three doses of linagliptin (1, 5 and 10 mg once daily) with placebo as add-on therapy to metformin treatment (ClinicalTrials.gov Identifier: NCT ). In an open-label treatment arm, patients were randomized to receive glimepiride (1, 2 or 3 mg once daily) administered as add-on therapy to metformin. These patients were advised to take a daily dose of 1 mg of glimepiride for 4 weeks. Thereafter, it was at the investigator s discretion to uptitrate the daily dose up to 3 mg. The study was conducted in 45 centres; 19 in the UK, 17 in Germany, five each in France, Slovakia and Ukraine, and four in Sweden. Eligible patients at screening who were already receiving metformin monotherapy entered a 2-week, open-label, run-in phase. Patients who were receiving metformin in combination with one other oral hypoglycaemic agent (other than rosiglitazone or pioglitazone) entered a 6-week period where the other oral hypoglycaemic agent was no longer administered, with an open-label placebo run-in phase for the last 2 weeks of this period. On completion of the 2-week run-in phase, patients with an HbA 1c level of % ( mmol mol) were randomized to one of the five treatment regimens in a 1:1:1:1:1 ratio and then underwent baseline assessments. After the 6-week wash-out period, and for the remainder of the study duration, patients who had previously received metformin in combination with one other oral hypoglycaemic agent (other than rosiglitazone or pioglitazone) did not receive further treatment with the oral hypoglycaemic agent they had received prior to enrolment. To maintain the blinded status of the treatments until randomization, sealed envelopes were created that contained the number of the randomized medication. The envelopes were numbered consecutively and opened by the investigator separately for each patient. The consecutive opening of the envelopes was regularly checked by the sponsor Diabetic Medicine ª 2010 Diabetes UK

3 Original article DIABETICMedicine Patients randomized into the double-blind arms took the trial medication once daily in the morning with 150 ml of water, within 30 min of finishing breakfast. Glimepiride was taken immediately before or during breakfast. Patients randomized to open-label glimepiride took a 1-mg dose for 4 weeks. After this, dosing was at the investigator s discretion. Patients continued the dose of metformin that they received at enrolment throughout the entire study. During the study, patients made visits to the study centre at weeks 2, 4, 8 and 12, with a follow-up visit at week 14. Patients underwent safety and efficacy assessments and were withdrawn if they had a fasting plasma glucose > 13.3 mmol l (measured on two separate days) at any visit; if they showed clinical signs of severe hypoglycaemia or a blood glucose level < 2.5 mmol l; if their dose of metformin changed; or if they received concomitant drugs that interfered with the study medication. This study was conducted and reported according to the principles of the International Committee for Harmonisation (ICH) Harmonised Tripartite Guideline for Good Clinical Practice. Adverse events were coded according to the Medical Dictionary for Regulatory Activities (MedDRA) coding system (version 10.0). All patients provided written informed consent to participate in the study and the protocol was approved by local ethics committees. Study assessments The primary endpoint of this study was the change in HbA 1c levels from baseline to week 12. Secondary endpoints included change from baseline in fasting plasma glucose concentrations after 12 weeks, and the number of patients achieving an HbA 1c value of 7.0% (53.0 mmol mol), as well as the number of patients achieving an HbA 1c lowering of at least 0.5% (5.5 mmol mol)after12weeksoftreatment. Other parameters of interest were inhibition of plasma DPP-4 activity, the proinsulin insulin ratio, the homeostasis model assessment (HOMA) indices for insulin resistance and insulin secretion, body weight, and plasma concentrations of linagliptin. The following were measured at each study visit: glycated haemoglobin (except week 2); fasting plasma glucose concentrations; adverse events; standard haematology and biochemistry safety laboratory assessments (except weeks 2 and 8); compliance and use of concomitant medication; vital signs; linagliptin plasma concentration; and plasma DPP-4 activity. Blood samples for linagliptin concentration and DPP-4 activity were collected pre-dose and at 1 h ( 0.5 h) and 2 h ( 1 h) after drug administration. Data were used for a population pharmacokinetic analysis, which is reported separately [22]. Body weight, electrocardiogram (ECG) and insulin proinsulin concentrations were measured at week 0 and week 12. Blood samples (1.2 ml) for the determination of HbA 1c levels were collected in ethylenediaminetetraacetic acid (EDTA) tubes, stored between 4 and 8 C and were analysed within 7 days. The percentage of HbA 1c from total haemoglobin was analysed by a validated high-performance liquid chromatography (HPLC) VariantÔ method in a National Glycohaemoglobin Standardization Programme Level I (NGSP)-certified central laboratory (MDS, Hamburg, Germany). Corresponding International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) standardized values were calculated using the relationship: IFCC value (in mmol mol) = (NGSP value )2.152) ) [23,24]. Blood (2.0 ml) was collected into sodium fluoride tubes for measurement of glucose concentrations, centrifuged immediately at 2500 g for 10 min at 4 C and the plasma supernatant collected in two aliquots (0.2 ml each) that were frozen until analysed by a validated enzymatic hexokinase method in a central laboratory (MDS, Hamburg, Germany). Proinsulin and insulin were measured in blood samples (2.7 ml) collected in EDTA-anticoagulant tubes and centrifuged within 30 min at 1000 g for 10 min at 2 8 C. The resulting plasma was stored at )20 C until analysed by validated immunoassays and chemiluminescence detection. For the determination of plasma DPP-4 activity and linagliptin plasma concentrations, blood samples (3 ml) were collected in EDTA tubes and analysed as described previously [14]. Adverse events were recorded throughout the study. The investigator recorded and assessed the severity of the adverse events and possible relationship to the study drug. Additional safety parameters included vital signs, safety laboratory assessments, a 12-lead electrocardiogram and pregnancy testing of female patients. All standard haematology and biochemistry safety assessments were performed by a central laboratory (MDS, Hamburg, Germany). Statistical methods The primary analysis was performed on the full analysis set, which followed the intent-to-treat principle as closely as possible. The full analysis set consisted of all randomized patients with at least baseline data and at least one adequate measurement of HbA 1c levels following at least 1 day of randomized treatment. A per-protocol set was created for sensitivity analysis. All safety analyses were performed on the treated set. The treated set included all patients who were dispensed study medication and had taken at least one dose of investigational treatment. The primary efficacy endpoint, the change in HbA 1c levels from baseline after 12 weeks of treatment, was analysed using analysis of covariance (ANCOVA) with baseline HbA 1c as a covariate. The potential superiority of treatment with linagliptin to placebo was tested by comparison of the changes in HbA 1c levels from baseline (tested sequentially from the highest to the lowest dose). A stepwise procedure reduced the risk of a type 1 statistical error. Each hypothesis test was performed at a = (one-sided). In addition, a stratified analysis was performed in an explorative way, which analysed the effect for subjects who took previous anti-diabetic medicine and therefore underwent the wash-out period, and for subjects who directly entered the run-in Diabetic Medicine ª 2010 Diabetes UK 1411

4 DIABETICMedicine Linagliptin added to metformin in Type 2 diabetes T. Forst et al. period. Treatment response for glimepiride was compared with placebo in an exploratory way to assess the sensitivity of the study to detect changes in the primary endpoint. The superiority hypothesis was similar to that described for the primary endpoint. Analysis of change in fasting plasma glucose concentration was performed in a similar manner to that described for HbA 1c levels, and was performed in an exploratory manner. If data were not available (e.g. for non-completers), last observation carried forward (LOCF) was used. In addition, a repeatedmeasurements approach was implemented as sensitivity analysis, which includes the data from the intermediate visits at weeks 4 and 8. All safety data were analysed using descriptive statistics (Table 4). The sample size calculation of this study was based on the assumption of a treatment effect of 0.5%, a standard deviation of the of 1.0% (10.9 mmol mol) and a (one-sided) test level of significance. A sample size per group of 73 randomized patients could detect a difference of the required magnitude at a significance level of 2.5% (one-sided) with 90% power. Thus, with five groups of 73 patients, at least 365 patients were required for this study. An interim analysis was introduced towards the end of the study as a fast-track analysis without impact on the conduct of the study. Based on recent data for the DPP-4 inhibitor class [25], the interim analysis was performed at the time when the primary endpoint of at least 225 patients was available. For the interim analysis, the study was unblinded for the sponsor, while investigators were kept blinded until the final lock of the database. At the time of unblinding the study for the interim analysis, all patients had performed all study procedures until week 8; there were 87 patients for which week 12 assessments were outstanding. Results Demographics and baseline characteristics Of the 669 patients enrolled in the study and screened, 333 patients were randomized and received linagliptin, placebo or glimepiride as add-on to existing metformin monotherapy (Fig. 1). At enrolment and screening, approximately 65% of the patients were receiving metformin monotherapy and 35% were being treated with metformin and one additional oral hypoglycaemic agent. Demographic and baseline characteristics Enrolled (n = 669) Randomized (n = 333) Excluded (n = 336) Reasons for exclusions: Violation of inclusion/exclusion criteria (n = 293, 87.2%) Withdrew informed consent (n = 27, 8.0%) Adverse events (n = 6, 1.8%) Lost to follow-up (n = 3, 0.9%) Other (n = 7, 2.1%) Placebo (n = 71) Linagliptin 1 mg (n = 65) Linagliptin 5 mg (n = 66) Linagliptin 10 mg (n = 66) Glimepiride (n = 65) Adverse event (n = 1) Lack of efficacy (n = 10) Non-compliant with protocol (n = 2) Withdrew consent (n = 1) Adverse event (n = 5) Lack of efficacy (n = 4) Withdrew consent (n = 4) Adverse event (n = 3) Lack of efficacy (n = 3) Lost to follow-up (n = 2) Withdrew consent (n = 2) Adverse event (n = 2) Lack of efficacy (n-3) Other (n = 1) Adverse event (n = 3) Lost to follow-up (n = 1) Completed (n = 57, 80.3%) Completed (n = 52, 80.0%) Completed (n = 56, 84.8%) Completed (n = 60, 90.9%) Completed (n = 61, 93.8%) FIGURE 1 Disposition of patients in the study and reasons for premature discontinuation of trial medication Diabetic Medicine ª 2010 Diabetes UK

5 Original article DIABETICMedicine Table 1 Baseline characteristics (randomized population) Characteristics Placebo Linagliptin 1mg Linagliptin 5mg Linagliptin 10 mg Glimepiride Total n Age (years), mean sd Male, n (%) 44 (62.0) 36 (55.4) 37 (56.1) 35 (53.0) 41 (63.1) 193 (58.0) Race, n (%) White 69 (97) 64 (99) 66 (100) 65 (99) 64 (99) 328 (99) Black 1 (1) (2) 0 2 (1) Asian 1 (1) 1 (2) (2) 3 (1) Body weight (kg), mean sd BMI (kg m 2 ), mean sd BMI group (kg m 2 ), n (%) < (28.2) 23 (35.4) 23 (34.8) 21 (31.8) 23 (35.4) 110 (33.0) (49.3) 22 (33.8) 24 (36.4) 28 (42.4) 27 (41.5) 136 (40.8) > (22.5) 20 (30.8) 19 (28.8) 17 (25.8) 15 (23.1) 87 (26.1) HbA 1c (%), mean sd HbA 1c (nmol mol), mean sd HbA 1c group, n (%) < 8.0%, <63.9 mmol mol 20 (28.6) 27 (42.2) 21 (33.9) 21 (31.8) 25 (39.1) 114 (35.0) %, mmol mol 35 (50.0) 26 (40.6) 26 (41.9) 31 (47.0) 28 (43.8) 146 (44.8) > 9.0%, >74.9 mmol mol 15 (21.4) 11 (17.2) 15 (24.2) 14 (21.2) 11 (17.2) 66 (20.2) FPG (mmol l), mean sd HOMA index for insulin resistance, mean sd HOMA index for insulin secretion, mean sd Plasma proinsulin insulin ratio, mean sd Disease duration (years), mean sd FPG, fasting plasma glucose; HOMA, homeostasis model assessment. of the randomized patients were distributed similarly between the treatment groups (Table 1). Overall, the patient group had a mean ( sd) baseline HbA 1c level of 8.3% ( 0.8%), 67.2 mmol mol ( 6.5); 80% of patients had a baseline HbA 1c level < 9.0% (74.9 mmol mol) and the mean ( sd) baseline fasting plasma glucose concentration was 10.3 mmol l( 2.3 mmol l). The mean duration of diabetes was 7.40 years ( years). At enrolment, the majority of the patients across the treatment groups were receiving a metformin dose of 1500 mg day (76.7%). Overall, 286 (85.9%) of patients completed the 12-week study. The proportion of patients discontinuing was similar across all treatment groups. The main reason for discontinuation was lack of efficacy and patients who discontinued for this reason occurred mostly in the placebo group. Rates of compliance (approximately 95%) with the study treatment were similar across the five treatment groups during the randomized phase. Pharmacokinetics and DPP-4 inhibition Mean plasma concentrations of linagliptin at trough remained relatively constant from week 4 to week 12 in all linagliptin dose groups. The mean plasma concentrations at weeks 4 and 12 values were: 3.79 and 3.57 nmol l for linagliptin 1 mg; 5.89 and 6.04 nmol l for linagliptin 5 mg; and 8.12 and 8.61 nmol lfor linagliptin 10 mg. The mean trough concentrations of linagliptin were in the range nmol l and nmol lforthe 5 and 10 mg doses, respectively. A less than 3-fold increase in trough concentration of linagliptin was achieved with the 10-mg dose compared with the 1-mg dose. The linagliptin 10-mg dose achieved approximately a 1.4-fold increase in trough levels compared with the 5-mg dose. The concentrations for the linagliptin 5- and 10-mg doses were sufficient to achieve DPP-4 inhibition of 80% at trough in 80% of the patients. At week 12, median ( sd) plasma DPP-4 inhibition in patients receiving linagliptin 5 and 10 mg once daily was 85% ( 13%) and 89% ( 15%) 24 h post-last linagliptin dose, respectively. In contrast, patients in the linagliptin 1-mg group achieved a median plasma DPP-4 inhibition of only 62% ( 20%). Efficacy on HbA 1c The addition of linagliptin (once daily) to ongoing metformin monotherapy resulted in a reduction in HbA 1c levels at week 12 compared with baseline for all three dose levels, whereas placebo administration was associated with a relative increase in HbA 1c from baseline (Fig. 2). The mean ( se) placebo-corrected reductions in HbA 1c at week 12 were )0.40% ( 0.14), )4.4 mmol mol ( 1.5) for linagliptin 1 mg (P = 0.006), )0.73% ( 0.14), )8.0 mmol mol ( 1.5) for linagliptin 5 mg (P < 0.001) and )0.67% ( 0.14), )7.3 mmol mol ( 1.5) for linagliptin 10 mg (P < 0.001). After adjustment for the number Diabetic Medicine ª 2010 Diabetes UK 1413

6 DIABETICMedicine Linagliptin added to metformin in Type 2 diabetes T. Forst et al. Table 2 Descriptive statistics of the from baseline at week 12 by subgroups (FAS-LOCF) Placebo Linagliptin 1 mg Linagliptin 5 mg Linagliptin 10 mg N [%], mean (SD) [mmol mol], mean (SD) N [%], mean (SD) [mmol mol], mean (SD) N [%], mean (SD) [mmol mol], mean (SD) N [%], mean (SD) [mmol mol], mean (SD) All Patients (0.74) 2.6 (8.0) 64 )0.14 (0.92) )1.5 (9.9) 62 )0.50 (0.81) )5.5 (8.9) 66 )0.42 (0.87) )4.6 (9.5) ATS No (0.62) 1.3 (6.7) 45 )0.33 (0.76) )3.6 (8.3) 37 )0.70 (0.62) )7.7 (6.8) 37 )0.64 (0.75) )7.0 (8.2) Yes (0.90) 5.2 (9.8) (1.11) 3.4 (12.2) 25 )0.19 (0.97) )2.1 (10.7) 29 )0.15 (0.94) )1.6 (10.0) HbA 1c baseline group <8.0% (0.45) 0.8 (5.1) (1.04) 1.6 (11.1) 21 )0.12 (0.80) )1.3 (8.7) 21 )0.39 (0.66) )4.3 (7.3) % (0.62) 5.6 (6.8) 26 )0.48 (0.61) )5.2 (6.2) 26 )0.58 (0.73) )6.3 (7.9) 31 )0.43 (0.72) )4.7 (7.9) >9.0% 15 )0.17 (1.05) )1.9 (11.7) 11 )0.05 (1.02) )0.5 (10.2) 15 )0.87 (0.80) )9.5 (8.7) 14 )0.46 (1.37) )5.0 (14.9) Gender Male (0.80) 3.4 (8.8) 36 )0.26 (0.95) )2.8 (10.2) 33 )0.42 (0.77) )4.6 (8.4) 35 )0.57 (0.81) )6.2 (8.8) Female (0.64) 1.4 (6.9) (0.87) 0.1 (8.7) 29 )0.58 (0.86) )6.3 (9.3) 31 )0.25 (0.91) )2.7 (9.8) Age Group <65 years (0.68) 3.1 (7.5) 40 )0.12 (0.96) )1.3 (10.4) 42 )0.49 (0.75) )5.4 (8.3) 37 )0.39 (0.89) )4.3 (9.8) 65 years (0.82) 2.1 (9.1) 24 )0.18 (0.86) )2.0 (9.6) 20 )0.51 (0.95) )5.6 (10.4) 29 )0.46 (0.85) )5.0 (9.2) BMI Group <30 kg m (0.77) 1.6 (8.2) 23 )0.26 (0.78) )2.8 (8.4) 22 )0.64 (0.63) )7.0 (6.9) 21 )0.36 (0.98) )3.9 (10.6) kg m (0.72) 2.8 (7.8) 22 )0.18 (0.85) )2.0 (9.4) 23 )0.30 (0.98) )3.3 (10.8) 28 )0.67 (0.85) )7.3 (9.3) >35 kg m (0.78) 3.4 (8.6) (1.14) 0.4 (11.4) 17 )0.58 (0.75) )6.3 (8.1) 17 )0.11 (0.64) )1.2 (7.0) Metformin Dose Group <1500 mg (0.79) 1.3 (8.6) 13 )0.25 (0.55) )2.7 (5.9) 11 )0.80 (0.69) )8.7 (7.5) 21 )0.49 (0.76) )5.4 (8.4) 1500 mg (0.73) 3.1 (8.1) 51 )0.11 (0.99) )1.2 (10.8) 51 )0.43 (0.83) )4.7 (9.1) 45 )0.40 (0.92) )4.4 (10.1) ATS, antidiabetic therapy status; FAS-LOCF, full analysis setªlast observation carried forward Diabetic Medicine ª 2010 Diabetes UK

7 Original article DIABETICMedicine Adjusted mean change in HbA1c (%) from baseline at Week Placebo (n = 70) * Linagliptin 1 mg (n = 64) ** Linagliptin 5 mg (n = 62) ** Linagliptin 10 mg (n = 66) Adjusted mean HbA 1c (SE) Baseline Week 4 Week 8 Week 12 FIGURE 3 Adjusted mean HbA 1c over time (full analysis set). Placebo ( ), 1 mg linagliptin (h), 5 mg linagliptin (s), 10 mg linagliptin ( ). FIGURE 2 Adjusted means for from baseline (%) at week 12 (full analysis set). Black bars are mean (+ se) change from baseline and grey bars are placebo-corrected change from baseline. All doses of linagliptin were significantly different from placebo in terms of HbA 1c reduction (*P < 0.01, **P < ). of pre-screening oral hypoglycaemic agents, the mean placebocorrected HbA 1c absolute changes from baseline were )0.39% ()4.3 mmol mol) (P = 0.005) in the linagliptin 1-mg group, )0.75% ()8.2 mmol mol) (P < 0.001) with the linagliptin 5-mg group and )0.73% ()8.0 mmol mol) (P < 0.001) with the linagliptin 10-mg group. The 95% confidence intervals for adjusted means from baseline at week 12 were ()0.66, )0.12) for linagliptin 1 mg, ()1.02, )0.48) for linagliptin 5 mg and ()0.99, )0.46) for linagliptin 10 mg. Changes in HbA 1c level over time are shown in Fig. 3,whichis based on a repeated-measurements analysis. In the placebo group, the HbA 1c levels increased slightly, but not significantly, from baseline up to week 8, and remained unchanged thereafter until week 12 (mean HbA 1c level: 8.48%). Reductions in HbA 1c with the linagliptin 5- and 10-mg doses followed each other closely and both achieved statistical significance from placebo (P < , both cases) by week 4, with the significant differences being maintained until the end of the study. In the linagliptin 1-mg dose group, the changes in HbA 1c levels were not as great, but statistically significant differences from placebo (P < ) emerged from week 8 onwards, with a mean maximum response at week 12 of 8.11% (65.1 mmol mol). The sensitivity of the study to detect differences between the treatment groups was confirmed by the glimepiride data. Treatment with the active comparator, glimepiride, for 12 weeks resulted in a decrease in HbA 1c levels from baseline of )0.68% ()7.4 mmol mol). A significant difference from placebo (P < ) in the change in HbA 1c levels from baseline to week 12 of )0.93% ()10.2 mmol mol) was observed for the glimepiride group [after adjustment for anti-diabetic therapy status, the reduction in HbA 1c levels was )0.90% ()9.8 mmol mol) (P < )]. There were no trends towards different responses to linagliptin with regard to age (P = 0.58), gender (P = 0.49) or BMI (P = 0.30); race was not evaluated as the majority of the patients were Caucasian. There was also no difference in response evident between the study countries (P = 0.47). The HbA 1c response depended on the anti-diabetic treatment status, whether the patients received another oral hypoglycaemic agent in addition to metformin before enrolment or not (Table 2). Across all treatments, including placebo, the results suggested that there were marked differences in the HbA 1c response depending on anti-diabetic therapy status (P < ). The magnitude of these differences was similar for all treatments, such that the placebo-corrected changes were comparable, and a formal test did not indicate an interaction between anti-diabetic treatment status and treatment (P = 0.86). A greater proportion of patients who received linagliptin (43.8% to 53.2%) showed reductions in HbA 1c from baseline of at least 0.5% (5.5 mmol mol) compared with placebo (12.9%). In addition, HbA 1c declined by 1% (10.9 mmol mol) or more in 27.4% and 22.7% of the patients treated with linagliptin 5 mg and 10 mg, respectively. In the placebo and linagliptin 1-mg groups, 7.1 and 14.1% of patients, respectively, showed HbA 1c reductions of at least 1% (10.9 mmol mol). Only one patient (1.4%) in the placebo group reached the absolute HbA 1c response criterion of 7.0% (53.0 mmol mol). In contrast, 10 patients (approximately 15%) who received linagliptin 1 mg, nine patients (approximately 15%) who received linagliptin 5 mg and 14 patients (approximately 21%) who received linagliptin 10 mg had HbA 1c levels 7% (53.0 mmol mol) after 12 weeks. Approximately 3% of the patients in the linagliptin 1- and 5-mg groups achieved HbA 1c levels 6.5% (47.5 mmol mol), whereas approximately 8% of the patients receiving linagliptin 10 mg recorded HbA 1c levels of 6.5% (47.5 mmol mol) at week 12. Efficacy on fasting plasma glucose Fasting plasma glucose reductions were significantly greater with all doses of linagliptin than with placebo at week 12 (see Fig. 4). The mean placebo-corrected fasting plasma glucose reductions for linagliptin were: )1.1 mmol l (P = ) for 1 mg; )1.9 mmol l (P < ) for 5 mg; and )1.6 mmol l (P < ) for 10 mg. After adjusting the mean fasting plasma glucose levels for oral hypoglycaemic agents received Diabetic Medicine ª 2010 Diabetes UK 1415

8 DIABETICMedicine Linagliptin added to metformin in Type 2 diabetes T. Forst et al. Adjusted mean fasting plasma glucose (FPG) (mmol/l) at Week Placebo (n = 68) before enrolment, the mean differences between linagliptin and placebo were: )1.1 mmol lfor1mg;)1.9 mmol l for 5 mg; and )1.7 mmol l for 10 mg. The 95% confidence intervals for the adjusted means change in fasting plasma glucose from baseline at week 12 were ()31, )7) for linagliptin 1 mg, ()47, )22) for linagliptin 5 mg and ()41, )17) for linagliptin 10 mg. Other outcomes 0.36 Linagliptin 1 mg (n = 63) Linagliptin 5 mg (n = 62) Linagliptin 10 mg (n = 66) FIGURE 4 Adjusted mean change from baseline in fasting plasma glucose (FPG) at week 12. Black bars are mean (+ se) change from baseline and grey bars are placebo-corrected change from baseline. All doses of linagliptin were significantly different from placebo in terms of FPG reduction (*P < 0.01, **P < ). After 12 weeks of treatment, a small change in mean body weight was observed in all linagliptin dosing groups: )0.15 kg (1 mg), )0.57 kg (5 mg) and )1.27 kg (10 mg). For patients treated with placebo, a mean weight loss of )0.84 kg was observed, whereas an increase of kg occurred in the glimepiridetreated group. Mean improvements from baseline in the HOMA index for insulin resistance were observed for all linagliptin doses: )0.8 for 1mg; )1.3 for 5 mg; and )1.3 for 10 mg. In contrast, the HOMA index increased by 0.5 from baseline after dosing with placebo. The mean values of the HOMA index for insulin * ** ** secretion following linagliptin increased by 3.4 for the 1-mg dose group, 8.1 for the 5-mg dose group and 7.9 for the 10-mg dose group, whereas it decreased by 2.5 in the placebo group (Table 3). Safety and tolerability The combination of linagliptin and metformin was well tolerated over 12 weeks. The proportion of patients reporting adverse events was similar across the linagliptin groups combined (44.2%), glimepiride (44.6%), and placebo (46.5%) groups. Sixteen patients discontinued from the study because of adverse events, with a similar proportion in each of the treatment groups. There was no apparent dose-dependent increase in reporting of adverse events with increasing linagliptin dose (Table 4). The most frequently reported adverse events were nasopharyngitis, diarrhoea and nausea, which were similarly distributed across all treatments. The most-reported clinical adverse events during the study were considered to be of mild or moderate intensity. However, 10 patients (3.7%) experienced adverse events that were judged to be of severe intensity defined by MedDRA as medically significant, but not life-threatening. The incidence of severe adverse events in the linagliptin 1-mg group was slightly higher (6.2%) than in the other treatment groups ( %). Ten patients experienced serious adverse events during the study, none of which was considered by the investigator to be related to the study medication. There were no vital sign or electrocardiogram safety issues for any of the treatment groups. Three patients experienced laboratory abnormalities that were judged to be a result of study medication; one of which (blood amylase increased by more than two times the upper limit of normal) occurred in the linagliptin 1-mg group. The other two abnormalities (decreased white blood cell count and hypoglycaemia) occurred in the glimepiride group. No episodes of hypoglycaemia were reported after dosing with linagliptin or placebo. Three (4.6%) subjects experienced hypoglycaemia after dosing with glimepiride. Table 3 Change from baseline in proinsulin insulin ratio and HOMA indices at week 12 (PPS-observed case) Placebo Linagliptin 1mg Linagliptin 5mg Linagliptin 10 mg Glimepiride Proinsulin insulin ratio Number of patients* Mean change from baseline (sd) 0.0 (0.1) )0.0 (0.1) )0.0 (0.1) )0.0 (0.1) )0.0 (0.0) HOMA index for insulin resistance Number of patients Mean change from baseline (sd) 0.5 (4.1) )0.8 (4.0) )1.3 (5.1) )1.3 (2.5) )1.7 (3.8) HOMA index for insulin secretion Number of patients Mean change from baseline (sd) )2.5 (31.5) 3.4 (23.5) 8.1 (34.5) 7.9 (19.5) )20.1 (210.7) *Per-protocol set (PPS): included all patients who had followed the protocol according to essential criteria. HOMA, homeostasis model assessment Diabetic Medicine ª 2010 Diabetes UK

9 Original article DIABETICMedicine Table 4 Adverse events with an incidence of 2% by treatment group (treated population) Placebo* n =71 n (%) Linagliptin* 1 mg n =65 n (%) Linagliptin* 5 mg n =66 n (%) Linagliptin* 10 mg n =66 n (%) Glimepiride n =65 n (%) Any adverse event 33 (47) 25 (39) 32 (49) 30 (46) 29 (45) Nasopharyngitis 7 (10) 4 (6) 5 (8) 5 (8) 4 (6) Upper respiratory tract infection 3 (4) 1 (2) (2) Diarrhoea 3 (4) 1 (2) 2 (3) 2 (3) 3 (5) Nausea 3 (4) 0 4 (6) 3 (5) 0 Increased blood glucose 2 (3) 2 (3) 0 1 (2) 1 (2) Arthralgia 2 (3) 2 (3) 1 (2) 0 0 Influenza 1 (1) 0 2 (3) 0 0 Urinary tract infection 1 (1) (3) 0 Musculoskeletal pain 1 (4) 0 2 (3) 1 (2) 0 Constipation (2) 2 (3) 1 (2) Dyspepsia (2) 1 (2) 2 (3) Increased triglycerides 0 2 (3) 0 1 (2) 0 Cough 0 1 (2) 2 (3) 1 (2) 0 Asthma 0 2 (3) (2) Fatigue 0 1 (2) 1 (2) 2 (3) 0 Hypoglycaemia (5) Any serious adverse event 1 (1) 3 (5) 1 (2) 4 (6) 1 (2) Discontinuation because of an adverse event 2 (3) 6 (9) 3 (5) 2 (3) 3 (5) *Double-blind study; open-label study. Discussion The present dose-response study assessed the efficacy, safety and tolerability of once-daily dosing with linagliptin (1, 5 and 10 mg) added to ongoing metformin therapy in patients with Type 2 diabetes and insufficient glycaemic control. Twelve weeks treatment with linagliptin resulted in clinically relevant reductions in HbA 1c and fasting plasma glucose compared with baseline for all linagliptin doses (1, 5 and 10 mg) and lower HbA 1c and fasting plasma glucose compared levels compared with placebo. It appears that maximum glucose-lowering effects were reached after 8 weeks, with a plateau thereafter. A greater proportion of patients receiving linagliptin treatment reached the study s predefined HbA 1c targets of reduction from baseline of 0.5% (5.5 mmol mol) and HbA 1c 7.0% (53 mmol mol). In this dose-response study, 5 mg linagliptin daily appeared to provide the maximum glycaemic effect, with the greatest percentage of patients achieving these targets (HbA 1c levels 0.5% (5.5 mmol mol): 53.2%; and 7.0% (53.0 mmol mol): 14.5%). Open-label treatment with glimepiride resulted in a numerically greater, yet not significantly different, reduction in HbA 1c compared with linagliptin. However, treatment with glimepiride was associated with a greater incidence of hypoglycaemia and an increase in body weight, both of which were not observed in patients treated with linagliptin. As this was an open-label comparison with only 12 weeks treatment duration, the long-term efficacy and safety profile of linagliptin vs. sulphonylurea treatment has to be assessed in longer-term studies. In this study, a greater proportion of patients demonstrating clinically meaningful changes in HbA 1c ( 0.5%, 5.5 mmol mol) appeared to be in those who had higher baseline HbA 1c levels ( 9.0%, 74.9 mmol mol). The effects reported with linagliptin are similar to those reported in a study where sitagliptin was added to existing metformin in a similar patient population [26]. The findings are consistent with the well-established concept of the indirect proportionality of degree in glycaemic control and the extent of glucose lowering produced by the oral hypoglycaemic agent [27]. Based on single- and multiple-dose studies in healthy volunteers and in patients with Type 2 diabetes, the therapeutic index of linagliptin is expected to be wide [14,15]. Previous studies indicate that the linagliptin dose levels used in the present study will provide true 24-h suppression of plasma DPP-4 activity at 5 and 10 mg dose levels and this was confirmed by pharmacokinetic and pharmacodynamic profiles of linagliptin in the current study population. At 12 weeks, trough levels of plasma linagliptin were sufficient to achieve 80% DPP-4 inhibition in 87% of patients in the 5-mg dose group. Increasing the linagliptin dose to 10 mg had only a minor effect on the overall DPP-4 inhibition profile, with 93% of patients achieving 80% inhibition at trough. The sustained 24-h action of linagliptin can be expected to complement the action of metformin. Three of the major pathophysiological defects of Type 2 diabetes are islet cell dysfunction, insulin resistance and increased hepatic glucose output [28]. Metformin decreases production of glucose by the liver and also targets insulin resistance by increasing the uptake and utilization of glucose. DPP-4 inhibitors such as linagliptin Diabetic Medicine ª 2010 Diabetes UK 1417

10 DIABETICMedicine Linagliptin added to metformin in Type 2 diabetes T. Forst et al. enhance the levels of the incretin hormones that increase insulin synthesis and release from pancreatic B-cells and lower glucagon secretion from pancreatic a-cells, leading to reduced production of glucose by the liver. Thus, when used together, metformin and linagliptin offer the potential to address the three key defects of Type 2 diabetes for improved glycaemic control. Indeed, linagliptin was more effective than placebo, in combination with metformin, in all measures of glycaemic control studied in this trial HbA 1c, fasting plasma glucose and parameters of B-cell function (HOMA indices). Similar results have been shown for sitagliptin combined with metformin in a longer (54-week) study [29]. Linagliptin was well tolerated at all doses without evidence for a dose-related reporting of adverse events with increasing linagliptin dose. The incidence of adverse events was comparable with that seen in the placebo group. There were no recorded hypoglycaemic episodes with linagliptin or placebo despite marked improvements in glycaemic control. This is consistent with the concept that incretin-based therapy should not cause hypoglycaemia [30]. Linagliptin did not cause an exacerbation of gastrointestinal signs or symptoms and thus offers an advantage over injectable GLP-1-based therapies. The latter are frequently associated with gastrointestinal side effects [31], possibly relating to the supraphysiological GLP-1 concentrations. A possible limitation of this study was the relatively short study duration and, thus, conclusions about longer-term efficacy and safety of linagliptin plus metformin must await the results of long-term trials that are currently ongoing. Conclusion In this study, linagliptin as add-on to metformin resulted in a statistically significant and, more importantly, clinically relevant improvement in glycaemic control in patients with Type 2 diabetes, without causing hypoglycaemia. Including additional covariates into the statistical model did not change the overall conclusion of the primary analysis. Of the linagliptin doses studied, 5 mg provided the greatest glycaemic efficacy. All doses of linagliptin were well tolerated and the incidence of adverse events was comparable with those found in placebo-treated patients. Competing interests BU-L, AR, UG-M, CF, KH, H-JW and KAD are employees of Boehringer Ingelheim, the sponsor of the study. Acknowledgements Data from this study have previously been published, in part, at the 69th Scientific Sessions of the American Diabetes Association, New Orleans, Louisiana, 5 9 June This study was supported by Boehringer Ingelheim Pharma GmbH and Co. KG. The authors would like to thank Dr Kristiane Wetzel, Dr Christine Rentzsch and Dr Barbara Withopf for bioanalytical support, and Sufian Chowdury and Brigitte Kölsch for programming support. Silvia Schuhholz provided data management services. Medical writing and editorial support was provided by Patrick Foley of PHASE II International Ltd, with the financial support of Boehringer Ingelheim. References 1 Bailey CJ, Turner RC. Metformin. N Engl J Med 1996; 334: Nathan DM, Buse JB, Davidson MB, Heine RJ, Holman RR, Sherwin R et al. Management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy: a consensus statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2006; 29: Inzucchi SE. Oral antihyperglycemic therapy for type 2 diabetes: scientific review. J Am Med Assoc 2002; 287: De Fronzo RA, Goodman AM. Efficacy of metformin in patients with non-insulin-dependent diabetes mellitus. The Multicenter Metformin Study Group. N Engl J Med 1995; 333: Campbell RK. Rationale for dipeptidyl peptidase 4 inhibitors: a new class of oral agents for the treatment of type 2 diabetes mellitus. Ann Pharmacother 2007; 41: Deacon CF, Ahren B, Holst JJ. Inhibitors of dipeptidyl peptidase IV: a novel approach for the prevention and treatment of type 2 diabetes? Expert Opin Investig Drugs 2004; 13: Langley AK, Suffoletta TJ, Jennings HR. Dipeptidyl peptidase IV inhibitors and the incretin system in type 2 diabetes mellitus. Pharmacotherapy 2007; 27: Holst JJ, Gromada J. Role of incretin hormones in the regulation of insulin secretion in diabetic and nondiabetic humans. Am J Physiol Endocrinol Metab 2004; 287: E199 E Drucker DJ. Enhancing incretin action for the treatment of type 2 diabetes. Diabetes Care 2003; 26: Elliott RM, Morgan LM, Tredger JA, Deacon S, Wright J, Marks V. Glucagon-like peptide-1 (7-36) amide and glucose-dependent insulinotropic polypeptide secretion in response to nutrient ingestion in man: acute postprandial and 24-h secretion patterns. J Endocrinol 1993; 138: Ahrén B. Emerging dipeptidyl peptidase-4 inhibitors for the treatment of diabetes. Expert Opin Emerg Drugs 2008; 13: Penfornis A, Borot S, Raccah D. Therapeutic approach of type 2 diabetes mellitus with GLP-1 based therapies. Diabetes Metab 2008; 34: S78 S Deacon CF, Holst JJ. Linagliptin, a xanthine-based dipeptidyl peptidase-4 inhibitor with an unusual profile for the treatment of type 2 diabetes. Expert Opin Investig Drugs 2010; 19: Hüttner S, Graefe-Mody EU, Withopf B, Ring A, Dugi KA. Safety, tolerability, pharmacokinetics, and pharmacodynamics of single oral doses of BI 1356, an inhibitor of dipeptidyl peptidase 4, in healthy male volunteers. J Clin Pharmacol 2008; 48: Heise T, Graefe-Mody EU, Hüttner S, Ring A, Trommeshauser D, Dugi KA. Pharmacokinetics, pharmacodynamics and tolerability of multiple oral doses of linagliptin, a dipeptidyl peptidase-4 inhibitor in male type 2 diabetes patients. Diabetes Obes Metab 2009; 11: Thomas L, Eckhardt M, Langkopf E, Tadayyon M, Himmelsbach F, Mark M. (R)-8-(3-amino-piperidin-1-yl)-7-but-2-ynyl-3-methyl- 1-(4-methyl-quinazolin-2-ylmethyl)-3,7-dihydro-purine-2,6-dione (BI 1356), a novel xanthine-based dipeptidyl peptidase 4 inhibitor, has 1418 Diabetic Medicine ª 2010 Diabetes UK

11 Original article DIABETICMedicine a superior potency and longer duration of action compared with other dipeptidyl peptidase-4 inhibitors. J Pharmacol Exp Ther 2008; 325: Eckhardt M, Hauel N, Himmelsbach F, Langkopf E, Nar H, Mark M et al. 3,5-Dihydro-imidazo[4,5-d]pyridazin-4-ones: a class of potent DPP-4 inhibitors. Bioorg Med Chem Lett 2008; 18: Blech S, Ludwig-Schwellinger E, Gräfe-Mody EU, Withopf B, Wagner K. The metabolism and disposition of the oral dipeptidyl peptidase-4 inhibitor, linagliptin, in humans. Drug Metab Dispos 2010; 38: Graefe-Mody U, Huettner S, Stähle H, Ring A, Dugi KA. Effect of linagliptin (BI 1356) on the steady-state pharmacokinetics of simvastatin. Int J Clin Pharmacol Ther 2010; 48: Eckhardt M, Langkopf E, Mark M, Tadayyon M, Thomas L, Nar H et al. 8-(3-(R)-aminopiperidin-1-yl)-7-but-2-ynyl-3-methyl-1-(4- methyl-quinazolin-2-ylmethyl)-3,7-dihydropurine-2,6-dione (BI 1356), a highly potent, selective, long-acting, and orally bioavailable DPP- 4 inhibitor for the treatment of type 2 diabetes. J Med Chem 2007; 50: Graefe-Mody EU, Padula S, Ring A, Withopf B, Dugi KA. Evaluation of the potential for steady-state pharmacokinetic and pharmaco dynamic interactions between the DPP-4 inhibitor linagliptin and metformin in healthy subjects. Curr Med Res Opin 2009; 25: Retlich S, Duval V, Graefe-Mody U, Jaehde U, Staab A. Impact of target-mediated drug disposition on linagliptin pharmacokinetics and DPP-4 inhibition in Type 2 diabetic patients. J Clin Pharmacol 2010; 50: Hoelzel W, Weykamp C, Jeppsson JO, Miedema K, Barr JR, Goodall I et al. IFCC reference system for measurement of haemoglobin A1c in human blood and the national standardization schemes in the United States, Japan, and Sweden: a method-comparison study. Clin Chem 2004; 50: Sacks DB, ADA EASD IDF Working Group of the HbA1c Assay. Global harmonization of hemoglobin A1c. Clin Chem 2005; 51: Richter B, Bandeira-Echtler E, Bergerhoff K, Lerch C. Emerging role of dipeptidyl peptidase-4 inhibitors in the management of type 2 diabetes. Vasc Health Risk Manag 2008; 4: Raz I, Chen Y, Wu M, Hussain S, Kaufman KD, Amatruda JM et al. Efficacy and safety of sitagliptin added to ongoing metformin therapy in patients with type 2 diabetes. Curr Med Res Opin 2008; 24: Bloomgarden ZT, Dodis R, Viscoli CM, Holmboe ES, Inzucchi SE. Lower baseline glycemia reduces apparent oral agent glucose-lowering efficacy. Diabetes Care 2006; 29: De Fronzo RA. Pathogenesis of type 2 diabetes mellitus. Med Clin North Am 2004; 88: Williams-Herman D, Johnson J, Teng R, Luo E, Davies MJ, Kaufman KD et al. Efficacy and safety of initial combination therapy with sitagliptin and metformin in patients with type 2 diabetes: a 54-week study. Curr Med Res Opin 2009; 25: Nauck MA, Kleine N, Orskov C, Holst JJ, Willms B, Creutzfeldt W. Normalisation of fasting hyperglycaemia by exogenous glucagon like peptide 1 (7-36 amide) in type 2 (non-insulin-dependent) diabetic patients. Diabetologia 1993; 36: Monami M, Marchionni N, Mannucci E. Glucagon-like peptide-1 receptor agonists in type 2 diabetes: a meta-analysis of randomized clinical trials. Eur J Endocrinol 2009; 160: Diabetic Medicine ª 2010 Diabetes UK 1419