Addition of Liraglutide to Patients With Type 2 Diabetes Treated With Multiple Daily Insulin Injections

Overview

Liraglutide, a GLP-1-analogue has been shown to be an effective treatment option in patients on oral anti-diabetes therapy with beneficial effects on both glycaemic control and weight. However, to date there are no clinical trials of liraglutide added to insulin therapy, a population of patients generally having worse glycaemic control and weight gain. In clinical guidelines, use of multiple daily insulin injections (MDI) is usually the final therapeutic option for type 2 diabetic patients. The primary study aim is to evaluate whether the addition of liraglutide, compared to placebo, reduces the HbA1c level for overweight and obese type 2 diabetes patients with inadequate glycaemic control treated with multiple daily insulin injections (MDI). MDI is defined as treatment with any basal insulin combined with separate meal time insulin injections before the main meals, i.e. an insulin regimen with premixed insulin is not considered as MDI. The planned study duration is 24 weeks and includes 120 patients at 15 centers in Sweden.

Background Patients with type 2 diabetes are generally treated with metformin and diet as first-line therapy (1). In Sweden sulphonylureas are generally recommended as the next treatment option due to their low cost and evidence for reducing diabetes complications (2). The next line of therapy includes adding basal or premixed insulin. As a final step multiple daily insulin injections with basal and prandial insulin have become the standard if glycaemic control does not meet targets. Obesity is another co-morbid condition in this population (3). The United Kingdom Prospective Diabetes Study (UKPDS) published 1998, the seminal trial demonstrating the importance of glycaemic control in type 2 diabetes, illustrated that insulin therapy (mostly basal insulin alone) is accompanied by significant weight gain in patients with type 2 diabetes (4). MDI generally results in even greater weight gain. Today there are few treatment options in patients with type 2 diabetes on MDI with poor glycaemic control. In patients with basal, intermediate or premixed insulin, addition of sitagliptin has shown further reduction in HbA1c of 0.6 percentage units (6 mmol/mol) with no change in weight (5). However, the effect in patients with the most severe disease, namely those requiring MDI, was not evaluated. In combination with metformin or sulphonylurea, adding the GLP-1-analogue liraglutide, compared with adding sitagliptin, has shown greater reductions in HbA1c and significant weight reductions (6). In an uncontrolled study in clinical practice, the addition of exanatide (n=22) or liraglutide (n=43) to patients on insulin therapy, the majority using MDI, was associated with a mean reduction in HbA1c by 1.0 percentage units (10 mmol/mol), reduction in weight of 7.1 kg and, simultaneously, a mean reduction in insulin dose by 37 units (7). At baseline patients were generally obese (mean BMI 36kg/m2), had poor glycaemic control, mean HbA1c of 8.9% (74 mmol/mol) and insulin doses of 92 units. The number of hypoglycaemic events was low, treatment satisfaction was high and patients rated hypoglycaemic episodes to be fewer than with insulin therapy alone. From these results it seems that there may be a beneficial effect associated with liraglutide when added to type 2 diabetic patients on MDI; however, this requires confirmation in a randomised clinical trial. This question is of particular concern since currently there are no treatment options for reducing weight and insulin doses while simultaneously improving glycaemic control for this patient group having most advanced disease. Further, many clinicians today use a therapeutic strategy consisting of MDI as a final treatment option, but many patients do still not reach target HbA1c and some even continue to have very poor glycaemic control (7). Therefore, the aim of this study is to study the effect on HbA1c when adding liraglutide to MDI in overweight and obese patients with impaired glycaemic control. The proposed study is a randomised double-blind trial for 24 weeks including patients with type 2 diabetes and poor glycaemic control despite being treated with MDI for at least 6 months. The aim is to study the effect on HbA1c when adding liraglutide to MDI in overweight or obese patients with impaired glycaemic control. Study drug preparation Novo Nordisk will provide study medication and treatment codes. Apoteket AB, Sweden, will handle study treatment logistics. Site monitoring Gothia Forum, Gothenburg, Sweden Data management Gothia Forum will provide the eCRF system to be used in the study. Data Management (including the randomisation system) and statistical analysis will be performed by Statistiska Konsultgruppen, Gothenburg, Sweden. Lab and continuous glucose monitoring All laboratory tests will be measured at the Research Centre for Laboratory Medicine at Karolinska University Hospital, Stockholm, Sweden. The DexCom G4 CGM device will be used for 21 days (7 days during the run-in period, 7 days during week 12 ± 7 days and 7 days during the last 2 weeks of the trial) for blinded monitoring of glucose values. Risk-benefit assessment Previously published observations from clinical practice (7) have shown that patients treated with multiple daily insulin injections (MDI) with concomitant addition of liraglutide (off-label) decreases the HbA1c by approximately 10 mmol/mo (1 percentage unit), reduces the weight by 7 kg and decreases the insulin dose with 37 units (7). These patients were prior to treatment overweight (BMI around 36), had HbA1c on average 74 mmol /mol and insulin doses were at 92 units. Patients also rated hypoglycaemias to be fewer when insulin was combined with liraglutide than when used alone. In general, patients reported an improved quality of life. There is always a risk of hypoglycaemia during insulin therapy irrespective of any additional glucose-lowering therapy. However, previous findings from clinical practice (7) and that the effect of GLP-1 on insulin release is only observed under hyperglycaemic conditions (8) indicate that liraglutide does probably not lead to major increased risks of hypoglycaemias. Since, liraglutide has been associated with a reduction in body weight in previous clinical trials (8) it is likely that such an effect exists also in individuals treated with insulin. In summary, the risk of side effects or adverse reactions with the combination of insulin and liraglutide is probably small in relation to the improved hyperglycaemic control and weight reduction that may exist when adding liraglutide in type 2 diabetic patients with basal and prandial insulin. Rational Multiple daily insulin injections (MDI) are regarded as the most advanced treatment regimen in patients with type 2 diabetes. At study start all patients have been treated with MDI for at least 6 months but are still suffering from poor glycaemic control. The patients included are also overweight or obese. It is therefore of importance to evaluate whether this patient group have a beneficial effect, on the glycaemic control and weight reduction, by adding liraglutide, since liraglutide has shown such beneficial effects in combination with oral glucose lowering drugs. Moreover, beneficial effects by liraglutide have been observed when used off-label in combination with insulin in clinical practice (7). Design This is a double-blind, placebo-controlled clinical trial with parallel-group design. The study involves 120 patients at 15 health centres in Sweden and includes 10 visits over 26 weeks. Treatments The study drug will be liraglutide which includes 3 different dosages: 0.6 mg, 1.2 mg and 1.8 mg. Liraglutide will be titrated to 1.8 mg by increasing the dose every week. Since subjects will have inappropriate glycaemic control (greater than or equal to 7.5%=58 mmol/mol HbA1c) no general reduction in insulin doses will be made, but if glucose curves during the run-in period show clearly lower mean glucose levels than those corresponding to HbA1c 7.5%=58 mmol/mol or are well-controlled before certain meals, insulin adjustments will temporarily be considered during the titration phase of liraglutide/placebo. Liraglutide will be available at a concentration of 6.0 mg/ml and supplied in a 3 ml disposable pen-injector. Placebo is supplied in an identical disposable pen-injector. Randomization After a run-in period of maximally 60 days, patients will be randomised to receive either 0.6 mg liraglutide or 0.6 mg placebo. Novo Nordisk will provide (coded) medication pens and the associated code-list. No study personnel other than a designated data manager responsible for setting-up randomisation will have access to the complete code-list until the study blind is broken. Apoteket AB/farmaci will handle study treatment logistics. Minimisation (optimal allocation) using a centralised web system (handled by Statistiska Konsultgruppen) will be used for randomisation. At randomisation each subject will be assigned a randomisation code that (blindly) represents the treatment (liraglutide or placebo) that the patient is randomised to. The patient will also be assigned a unique and anonymous Subject ID. Breaking blind The study is double-blinded. The randomisation codes will not be broken until last subject's last visit and when all study data are collected, the study database is cleaned and declared locked. Apoteket AB/farmaci will be responsible for keeping the complete randomisation list. The code for a particular subject may be broken in a medical emergency situation. Apoteket/farmaci is equipped with 24 hour emergency telephone service in case of need for breaking blind in an individual patient. Telephone number: Weekdays 08.30-16.00: 040-33 26 74. Any other time: 040-331000 - ask for emergency telephone service. The subject should be withdrawn from the study in case of a code break but be included in the safety population in the statistical analyses. Duration The expected duration of subject participation is approximately 32 weeks (maximally 60 days run-in, and 24 weeks follow-up after randomisation). Study population The study will include 120 patients at 15 centres in Sweden. Patients requiring MDI and who are on no other diabetes drugs except metformin will be recruited and included in the study after receiving oral and written information about the study and given informed consent and fulfilling all the inclusion criteria and none of the exclusion criteria. The study population will be identified among diabetic outpatient clinics at hospitals and will be recruited on a community-wide basis and in communication with primary care physicians. The written information will include objects and purpose of the study, risks and benefits, and a detailed description of the events during the study. The patient will have the opportunity to ask additional questions at the first clinic visit The patients will be randomised 1:1 to liraglutide or placebo Randomisation codes will be assigned strictly sequentially as subjects are eligible for randomisation. If a subject discontinues participation in the study, his/her enrolment/subject number cannot be reused. Drop-outs (expected to be \~5%) will not be replaced. Study Stop Criteria The Sponsor/Investigator may decide to stop the trial or part of the trial at any time for medical reasons. Furthermore, the manufacturers of the study drugs may revoke their products, which may lead to the entire study may be interrupted. If a trial is prematurely terminated or suspended, the Investigator should promptly inform the patients and ensure appropriate therapy and follow-up. Furthermore, the Investigator should promptly inform the Ethics committee and provide a detailed written explanation. The regulatory authority should be informed according to national regulations. Drop-outs Patients who drop-out from the study or receive rescue therapy shall be requested for follow-up on an additional visit. HbA1c, FPG, C-peptide, pro-insulin, amylase, lipase, liver transaminases, creatinine, calcitonin, fasting blood lipids, adiponectin, sensitive CRP, apolipoproteins, weight, blood pressure, insulin dose, waist-hip ratio, abdominal sagittal diameter, concomitant medications and treatment satisfaction shall be checked. Blood samples for biobank will be taken. Insulin doses and other medications recorded. If drop-out exists the reason shall be recorded. Regarding rescue therapy the variables above shall be checked if possible before rescue therapy is given. Investigational product and treatment of subjects The trial product will be dispensed to each subject as required according to treatment group. The centralized web randomisation system will allocate trial product Dispensing Unit Number (DUN) to the subject at the randomisation visit. The correct DUN must be dispensed to the subject. The Investigator must ensure availability of proper storage conditions and record and evaluate the temperature (at least every working day). Storage facilities should be checked frequently. A log to document the temperature must be kept. The study is planned for 24 weeks and in addition a run-in period of maximally 60 days. Capillary glucose values will be measured before each meal and 1.5-2 hours after meals at 2 days during the 60-day run-in period, 2 days per week during the first 3 weeks of the trial and 2 days during the last 2 weeks of the trial. Capillary glucose curves performed during the run-in period and the last two weeks of the study will always be measured after the week with CGM to not disturb the interpretation of CGM data. The Contour XT SMBG meter (Bayer) shall be used for capillary glucose measurements during the study. The SMBG meter shall be checked at eash visit before initiating CGM, i.e. during the run-in period, week 12 (+- 7 days) and week 23-24. Calibration will be performed according to the manufacturer's standards. The SMBG meter shall be introduced at visit 4. During 7 days of the run-in period, 7 days of week 12 ± 7 days and 7 days of the 2 last weeks in the study, CGM with DexCom 4G will be used. CGM should be taken for 7 days, but a minimum of 5 days is acceptable. If a shorter period exists, a new sensor shall be set to get data for a total of 7 days (A total period of 5 days is acceptable). Investigators will recommend patients measure 3-4 glucose values per day before meals and at bedtime daily according to clinical guidelines for potentially adjusting insulin doses, diet or physical activity. Liraglutide/placebo will be administered at 0.6 mg the first week, 1.2 mg the second week and 1.8 mg the third week and onward. Dose increase periods can be extended based on subject's tolerance to the trial product. If the subject does not tolerate 1.8 mg liraglutide/placebo due to side effects, the dose shall be reduced to 1.2 mg liraglutide/placebo. Correspondingly, if 1.2 mg liraglutide/placebo is not tolerated 0.6 mg liraglutide/placebo shall be used. Since subjects will have inappropriate glycaemic control (greater than or equal to 7.5%=58 mmol/mol in HbA1c) no general reduction in insulin doses will be made. However, glucose profiles during the run-in period shall be judged by the clinician. In the following cases the clinician will consider to take the following actions when liraglutide is added: A) If the patient has a mean glucose level lower than 8.5 mmol/l the clinician should consider reducing the total insulin dose by 10%-20% when adding liraglutide 0.6 mg B) If glucose levels before a certain meal are well-controlled (less than 7 mmol/l) the clinician should consider reducing the meal-time insulin dose administered at the previous meal by 10%-20% when adding liraglutide 0.6 mg. If daytime hypoglycaemia has occurred a larger than 20% reduction in the insulin dose at the previous meal shall be considered when adding liraglutide 0.6 mg. C) If fasting glucose is 5.0-7.0 mmol/l the basal insulin shall be reduced by 20%-30% when adding liraglutide 0.6 mg. D) If fasting glucose is less than 5 mmol/l or nocturnal hypoglycaemia has occurred during the run-in period basal insulin shall be reduced by 20%-40%. Novel evaluations with glucose measurements during 4-5 days shall be performed to confirm that FPG is greater than 5 mmol/l and no new nocturnal hypoglycaemia exists before adding liraglutide. A novel evaluation according to point A)-D) shall then be performed from the new glucose values when adding liraglutide. When increasing liraglutide from 0.6 mg to 1.2 mg as well as from 1.2 mg to 1.8 mg the corresponding actions as points A-D above shall be considered based on glucose values and hypoglycaemias during the previous week. After titrating liraglutide/placebo to 1.8 mg in patients where insulin doses at any stage have been reduced, the insulin doses will be increased until the original doses are reached. However, the titration will be stopped, i.e., the insulin doses will not be titrated to the original doses, if glucose levels at any stage in the phase of increasing doses are already on target. For example, if the mean glucose level before a certain meal is less than 7 mmol/l after adding liraglutide/placebo 1.8 mg, the insulin dose before the previous meal will not be increased although the dose has been reduced. Correspondingly, if mean FPG is less than 7.0 mmol/l the basal insulin dose will not be increased although it has been reduced. The total insulin dose shall not be titrated higher than the original total insulin dose at any point during the study, except if rescue therapy is required. The diabetes educator or physician will have regular telephone contact with the patient during the initial period when liraglutide is titrated and for possibly adjusting insulin doses. At the 6, 12 and 18 week follow-up visits and other contacts with the patient, e.g. telephone contacts, the insulin doses can be increased in patients with a lower insulin dose than the original insulin dose before randomisation so long as glucose levels then not are on target (less than 7.0 mmol/l fasting or before meal). However, the insulin doses shall not be increased higher than the original doses, except if rescue therapy is undertaken In daily activity, the patient will continue to adjust meal-time insulin doses in the same way as previously performed in clinical practice. Rescue Criteria If the fasting self-measured plasma glucose (fasting SMPG) values taken on three separate days or if any of the FPG samples analysed by the central laboratory exceeds the limit of 15.5 mmol/l from baseline to week 12 or 13.5 mmol/l from week 12 to week 24, the subject should be called for an unscheduled visit as soon as possible. The next scheduled visit should not be awaited. A confirmatory FPG should be obtained and analysed by the central laboratory. If this FPG exceeds the above described values, and no treatable intercurrent cause for the hyperglycaemia has been diagnosed (e.g. compliance to liraglutide/placebo, missing of insulin dose in the evening, infection), the subject should receive rescue therapy. Rescue therapy is defined as an increase of the total insulin dose above the original total insulin dose (i.e. total insulin dose at baseline, at randomisation). It should be noticed that rescue therapy is NOT given if the total insulin dose was initially reduced in the trial and the total insulin dose is increased by the physician due to FPG above thresholds but not higher than the total original insulin dose. Rescue therapy will not exist either if a patient increases the total insulin dose at any point in time above the total original insulin dose by, e.g., self-adjustments of meal-time insulin doses in daily activity. Other medication, which is considered necessary for the subject's safety and well-being, may be given at the discretion of the investigator. Medications prescribed before study entry should preferably remain unchanged during the study period unless changes are made for safety reasons. Procedures for Monitoring Subject Compliance Compliance and adherence to protocol will be checked at each contact with the patient. Subject compliance will be assessed by DMG monitoring accountability. The unused amount of trial product will be assessed against the dispensed amount and, in case of discrepancies, the subjects must be asked about compliance. The individual delegated by the Investigator must keep track of all received, used, partly used and unused trial products and, if possible, all empty packaging. This shall be properly documented. Used, partly used and unused trial products and empty packaging must be stored separately from non-allocated trial products. Used, partly used and unused trial products must be stored until the monitor has performed drug accountability. Destruction of trial products will be performed in accordance with local legislation and recorded on a destruction form. Liraglutide/Placebo Not in use: The liraglutide/placebo pen-injector must be stored in a refrigerator at a temperature between +2°C and +8°C. Keep away from the cooling element. Do not freeze and do not use if it has been frozen. In use: After first opening the liraglutide/placebo pen-injector can be stored for one month at temperatures below +30°C or in a refrigerator between +2°C and +8°C. The liraglutide/placebo pen-injector must be protected from all sources of light, and the pen cap should be kept on when the pen is not in use. The liraglutide/placebo should not be used if it does not appear clear and colorless. In case of incorrect storage the Investigator or site staff must contact Novo Nordisk without delay. Trial product(s) must be set on-hold and not dispensed to subjects until notified by Novo Nordisk. The individual delegated by the Investigator must keep track of all received, used, partly used and unused trial products and, if possible, all empty packaging, by using the drug accountability module in the centralised web system. No trial product should be dispensed to any person not enrolled in the trial. Returned trial product(s) (used, partly used or unused and if possible, empty packaging material) must be stored separately from non-allocated trial product(s). Destruction of trial products will be done according to local laws and will be recorded on a destruction form, which must be signed by the individual responsible for destruction. Standardised procedures for administration of questionnaires The questionnaires will be filled in at the study site. The patients should be allowed to sit alone in a reasonable quiet environment to answer the questions. It will be emphasized that patients complete the questionnaires prior to clinical measurements and before meeting a doctor. Questionnaires should be answered by the patient herself alone, however, the nurse/assistant will be informed about helping the patients to complete the questionnaires if necessary, however without influencing the patients' responses. Ensure the patient confidentiality. Study nurse/assistant should check the questionnaires for completeness. The principal investigator will assure that appropriate training relevant to the study is given Trial Procedures Run-in period The purpose of the run-in period is to check glucose values as a basis for possible insulin adjustments and to perform CGM before randomisation. Blinded CGM will be performed during 1 week of the 60-day run-in period. Capillary glucose values before meals and 1.5 hours after meals will be measured during at least 2 days, after the CGM-period. Treatment satisfaction (DTSQs) will be checked, preferentially before initiating CGM, since CGM may influence treatment satisfaction. Randomization visit Glucose profiles and hypoglycaemias during the run-in period will be checked to judge whether insulin reductions shall be performed (See section 8.2 Titration Algorithm). At the randomisation visit the following variables will be measured before randomisation and blood samples will be sent to the Research Centre for Laboratory Medicine at Karolinska University Hospital. All blood samples are recommended to be fasting to facilitate logistics since several tests require fasting conditions. * Systolic and diastolic blood pressure after at least 5 minutes rest in sitting position with appropriate cuff-size * HbA1c and fasting glucose * Amylase, lipase, creatinine, liver transaminases * Fasting Blood Lipids (Triglycerides, Total-, HDL- and LDL-Cholesterol), fasting apolipoproteins * Weight * Waist-hip ratio and abdominal sagittal diameter * Type of insulin and doses of insulin and metformin * AE, SAE and hypoglycaemia * Adiponectin, sensitive CRP * GAD-antibodies, calcitonin, amylase, lipase * Fasting C-peptide and fasting pro-insulin * Biobank samples * Type of insulin, doses of insulin and doses of metformin recorded. * Other types of medications recorded * A morning urin sample of Albuminuria/Creatinine ratio will be measured with local laboratory analysis either at inclusion/exclusion visit or at randomisation visit. Web randomisation will be performed with minimisation (optimal allocation) to liraglutide or placebo in 1:1 proportion. Telephone contacts The diabetes educator or the physician will have telephone contact with the patient as follows: * The day after starting liraglutide/placebo * One of the 2 days before increasing liraglutide/placebo to 1.2 mg * The day after the increase to 1.2 mg * One of the two days before increasing liraglutide/placebo to 1.8 mg * The day after increasing liraglutide to 1.8 mg * One week after the increase of liraglutide/placebo to 1.8 mg More frequent telephone contacts during the initial weeks or later in the study will be made depending on the decision of the diabetic educator or physician if needed for the individual patient. At each telephone contact, insulin adjustments will be considered according to specified algorithms, i.e.: * if hypoglycaemia exists * If glucose values are well-controlled before titrating liraglutide/placebo In patients where insulin doses are reduced at any stage during the titration of liraglutide, regular telephone contacts every 3-4 day after titrating liraglutide/placebo to 1.8 mg shall be performed. Insulin doses will then be increased until the original insulin doses are reached. The increase of insulin doses will be stopped at a lower level earlier before original doses are reached if glucose levels already are on target. Clinical Visits Clinical visits will be performed at weeks 6, 12, 18 and 24. Visit at Week 6, 12 and 18 At the 6 weeks follow-up visit (± 7 days), 12 weeks follow-up visit (± 7 days) and 18 weeks (± 7 days) follow-up visit the following variables will be measured: * Systolic and diastolic blood pressure after at least 5 minutes rest in sitting position with appropriate cuff-size * HbA1c, Fasting plasma glucose sent to central laboratory * Weight * Type of insulin and doses of insulin and doses of metformin * AE, SAE and hypoglycaemia * Changes to concomitant medication * Trial product dispensing * At week 12: blood lipids, lipase, amylase, liver transaminases, creatinine, calcitonin and biobank samples sent to central laboratory. Waist-hip ratio and abdominal sagittal diameter shall be measured. Week 12 ± 7 days * Blinded CGM will be performed during 1 week. * Treatment satisfaction (DTSQs) will be evaluated, preferentially before CGM since CGM may influence the treatment satisfaction. Two last weeks (week 23-24) * Blinded CGM will be performed during 7 days. Capillary glucose values measured before and 1.5 hours after meals during 2 days. * Treatment satisfaction (DTSQs and DTSQc) will be performed, preferentially before CGM since CGM may influence treatment satisfaction. Week 24 At the final 24 weeks follow-up visit (± 7 days) the following variables will be measured (all blood samples sent to central laboratory): * Physical investigation * Systolic and diastolic blood pressure after at least 5 minutes rest in sitting position * HbA1c and fasting plasma glucose * Fasting blood lipids (Triglycerides, Total-, HDL- and LDL-Cholesterol) and apolipoproteins * Weight * Waist-hip ratio and abdominal sagittal diameter * Sensitive CRP, adiponectin, liver transaminases, creatinine, calcitonin, lipase and amylase * Type of insulin and doses of insulin and metformin * AE, SAE and hypoglycaemia * Changes to concomitant medication * Fasting C-peptide and fasting pro-insulin * Biobank samples * A morning urin sample of Albuminuria/Creatinine ratio will be measured with local laboratory analysis. Hypoglycaemias Hypoglycaemias will be defined according to the general guidelines from the American Diabetes Association of definitions of hypoglycaemias in study settings (9). Hypoglycaemias will be divided into documented symptomatic hypoglycaemias, documented asymptomatic hypoglycaemias and severe hypoglycaemias. Patients will be counselled of typical symptoms of hypoglycaemia and given a diabetes diary for recordings of glucose values, signs of hypoglycaemias and recovery from symptoms after intake of carbohydrates. The frequency of hypoglycaemia with glucose values below 3.0 mmol/l and 4.0 mmol/l will both be used as endpoints. Severe hypoglycaemias will also be recorded in the diabetes diary. If a severe hypoglycaemia appears the responsible physician for the patient should be informed as soon as possible. The following parameters of hypoglycaemia will be recorded in accordance with guidelines for study design of the American Diabetes Association: * Number of documented symptomatic hypoglycaemias (PG less than 4.0 mmol/l). * Number of documented asymptomatic hypoglycaemia (PG less than 4.0 mmol/l). * The number of symptomatic and asymptomatic documented hypoglycaemia with PG less than 3.0 mmol/l will also be recorded. * Number of glucose values recorded by CGM below 3.0 mmol/l and 4.0 mmol/l respectively. * Number of severe hypoglycaemias. Patients shall be informed regarding the symptoms of hypoglycaemia and requested to immediately perform a finger stick glucose measurement if symptoms occur that may be related to hypoglycaemia, but to avoid delay in treating these symptoms. Full analysis set (ITT-population) The full analysis set (ITT-population) consists of all randomised patients who received at least one dose of study medication and has at least one follow-up measurement. Primary Efficacy Analysis Primary efficacy analysis will be change in HbA1c from baseline to 24 weeks follow-up between the two treatment groups using analysis of covariance (ANCOVA) with HbA1c at baseline as covariate on the full analysis set (ITT-population), two-sided test on significance level 0.05. If HbA1c from 24 weeks follow-up is missing the last observation carry forward (LOCF) principle from 6, 12 and 18 weeks will be applied. All measurements obtained after rescue therapy should be excluded in all efficacy analyses.