A Randomized Controlled Trial of Liraglutide Versus

Therapeutics

The Journal of Clinical Pharmacology A Randomized Controlled Trial of Liraglutide 2015, XX(XX) 1–8 © 2015, The American College of Clinical Pharmacology Versus Insulin Detemir Plus Sitagliptin: Effective DOI: 10.1002/jcph.483 Switch From Intensive Insulin Therapy to the Once-Daily Injection in Patients With Well-Controlled Type 2 Diabetes

Yuichiro Inoue, MD1, Akinobu Nakamura, MD, PhD1,2, Yoshinobu Kondo, MD1,3, Kumiko Hamano, MD4, Shinobu Satoh, MD, PhD3, and Yasuo Terauchi, MD, PhD1

Abstract This study aimed to compare the efficacy and safety of liraglutide versus insulin detemir plus sitagliptin in Japanese patients with type 2 diabetes treated with a basal-bolus insulin regimen. In this multicenter, open-label trial, 90 patients whose diabetes had been controlled well or moderately (glycated hemoglobin [HbA1c] 7.3%) with basal-bolus insulin regimen were randomly assigned to a liraglutide group or a detemir group and were followed up for 24 weeks. The primary end point was HbA1c change from baseline to 24 weeks. Of the 90 enrolled patients, 82 completed this trial. At 24 weeks, the mean changes in HbA1c from baseline were 0.1% Æ 0.9% versus 0.3% Æ 0.8% in the liraglutide versus detemir groups, respectively (P ¼ .46). The “overall” satisfaction score for the Diabetes Treatment Satisfaction Questionnaire changed from 25.2 Æ 7.4 to 29.9 Æ 5.3 (P < .001) and from 26.4 Æ 6.1 to 28.3 Æ 6.4 (P ¼ .12) in the liraglutide and detemir groups, respectively. Although the mean change difference in HbA1c between both groups was not significant, switching from a basal-bolus insulin regimen to liraglutide once daily improved patient satisfaction levels without loss of glycemic control.

Keywords liraglutide, basal-bolus insulin regimen, patient’s satisfaction, type 2 diabetes

In patients with type 2 diabetes and uncontrolled ing from frequent injections to once-daily injections may hyperglycemia, a basal-bolus insulin regimen is the be convenient for these patients and may improve their recommended treatment compared with biphasic, prandi- treatment satisfaction. al, or basal insulin regimens.1,2Tight glycemic control is A new class of incretin-based antidiabetic agents, associated with reduced risks of long-term microvascular glucagon-like peptide-1 (GLP-1) receptor agonists and complications for these patients.3–5 However, intensive glucose-lowering therapy can result in an increased frequency of hypoglycemia and weight gain.6,7 Further- more, the intensive treatment arm of the Action to Control 1Department of Endocrinology and Metabolism, Graduate School of Cardiovascular Risk in Diabetes study, targeting glycated Medicine, Yokohama City University, Yokohama, Japan 2Division of Immunology and Metabolism, Department of Internal hemoglobin (HbA1c) < 6.0% (42 mmol/mol), was discontinued because of higher mortality and cardiovas- Medicine II, Graduate School of Medicine, Hokkaido University, Sapporo, Japan cular events in this group compared with the standard 3Department of Endocrinology and Metabolism, Chigasaki Municipal therapy group targeting HbA1c from 7.0% (53 mmol/mol) Hospital, Chigasaki, Kanagawa, Japan 8 to 7.9% (63 mmol/mol). 4Department of Diabetes and Endocrinology, Kanto Rosai Hospital, The prevalence of type 2 diabetes mellitus in the Kawasaki, Kanagawa, Japan elderly population is increasing, and diabetes is associated Submitted for publication 24 November 2014; accepted 11 February with an increased incidence of functional disability. The 2015. causes are multifactorial, including microvascular and macrovascular complications, comorbidities, and aging.9 Corresponding Author: In addition, diabetes has been associated with a greater Yasuo Terauchi, MD, PhD, Professor, Department of Endocrinology 10,11 and Metabolism, Graduate School of Medicine, Yokohama City risk of cognitive decline and function. Therefore, it is University, 3-9 Fukuura, Kanazawa-ku, Yokohama 236-0004, Japan assumed that there will be increased difficulty among this Email: [email protected] population to continue frequent insulin therapy. Switch- HbA1c is expressed by NGSP values. 2 The Journal of Clinical Pharmacology / Vol XX No XX 2015 dipeptidyl peptidase–4 (DPP-4) inhibitors, is available for was continued during the first week, reduced by half the treatment of type 2 diabetes. The GLP-1 receptor during the second week, and discontinued at the end of the agonist, especially liraglutide, improves postprandial second week. At 4 and 8 weeks, if self-measured fasting glycemic control and reduces body weight without blood glucose (FBG) was >160 mg/dL, glimepiride increasing the risk of severe hypoglycemia in patients 0.5 mg/day was added, and the dose was increased to with type 2 diabetes.12,13 Adding a DPP-4 inhibitor to 1.0 mg/day, if necessary. insulin therapy is also useful in glycemic control without In the detemir group, the starting dose was used causing severe hypoglycemia, especially for long-acting throughout the study. At 0 weeks, sitagliptin (50 mg/day, and once-daily insulin regimens.14 after breakfast) was added to insulin detemir. During the However, there are no prospective studies assessing first 8 weeks, insulin detemir was titrated once every 2 the effectiveness of switching from a basal-bolus insulin weeks by patients, based on self-measured FBG, aiming regimen to once-daily injection. Therefore, in this study, for a target value of 90 FBG 130 mg/dL. we compared the efficacy, safety, and patient satisfaction Subsequently, the insulin detemir dose was fixed for 16 with liraglutide versus insulin detemir plus sitagliptin. weeks. Liraglutide (6.0 mg/mL) and insulin detemir (100 units/mL) were injected subcutaneously once daily with injection pens. Patients and Methods Trial Design Outcomes The study was a 24-week open-label, randomized, The primary outcome measure was the change in HbA1c controlled trial conducted in Japanese patients with type from randomization (week 0) to week 24 after the 2 diabetes at 3 institutions (Yokohama City University administration of liraglutide or insulin detemir. The School of Medicine, Yokohama, Japan; Chigasaki Munic- secondary outcome measures were the changes in FPG, ipal Hospital, Chigasaki, Japan; and Kanto Rosai Hospital, self-monitoring blood glucose (SMBG), and body weight Kawasaki, Japan). The trial was registered at the University at 0, 8, 16, and 24 weeks. Our patients measured blood Hospital Medical Information Network (UMIN) Center glucose 6 times per day (each before and 2 hours after the under the identifier UMIN 000007009. The study protocol start of meals), once weekly. We calculated the mean was compliant with the Declaration of Helsinki, and it was postprandial blood glucose based on the data measured at approved by the ethics committee of the participating sites. the latest 2 times. All patients provided their written informed consent. Eight Adverse effects, such as hypoglycemia and gastrointes- weeks before randomization, the long-acting insulin tinal symptoms, were monitored during this trial. Hypo- injection was switched to insulin detemir once daily glycemia was defined as an FPG of <70 mg/dL or the before breakfast, and the oral glucose-lowering agents presence of hypoglycemic symptoms. The occurrence of were discontinued. Patients were randomized in a 1:1 ratio other symptoms was diagnosed by the physicians in to liraglutide or insulin detemir group by the institution charge. using central computer-based randomization. Patient-reported outcomes were measured using the The 90 randomized patients with type 2 diabetes were Diabetes Treatment Satisfaction Questionnaire, status aged 20 years and were treated with a basal-bolus version (DTSQs). 15 The DTSQs includes 8 items, and insulin regimen for at least 3 months before screening. responses are scored on a 7-point scale, from þ6 to 0. The Inclusion criteria included a diagnosis of type 2 diabetes, scores of the 6 items of the DTSQs (current treatment, HbA1c level 7.3% (56 mmol/mol), and treatment with a convenience, flexibility, understanding, likelihood of basal-bolus insulin regimen and insulin dose of 30 units. recommending their present treatment [recommend], Patients with a history of diabetic ketoacidosis or diabetic and satisfaction to continue with their present treatment coma within 6 months prior to the study entry, who were [continue]) were summed as the “overall” treatment pregnant or lactating, with severe infection or trauma, satisfaction score, ranging from þ36 to 0, with higher undergoing surgical operation during the study, undergo- scores denoting greater treatment satisfaction. In addition, ing steroid therapy, with a history of cardiac failure, with we included “perceived frequency of hyperglycemia severe organ dysfunction, with type 1 diabetes, with (Hyperglycemia)” and “perceived frequency of hypogly- insulin secretion disorders (fasting serum C-peptide < cemia (Hypoglycemia)” in the DTSQs, which were rated 0.5 ng/mL), and with a history of hypersensitivity to on a scale of þ6 (“most of the time”) to 0 (“never”). liraglutide, detemir, or sitagliptin were excluded. Patients completed a Japanese version of the DTSQ at 0 In the liraglutide group, patients underwent a 2-week and 24 weeks.16 dose escalation, starting at 0.3 mg once daily with weekly increments of 0.3 mg, reaching a final daily dose of 0.9 mg Statistical Analysis by the end of the second week. Subsequently, the A total of 74 patients (37 per treatment) were required to liraglutide dose was fixed for 22 weeks. Insulin detemir provide 80% power to detect a difference between Inoue et al 3

Table 1. Baseline Characteristics of the Subjects

Total (n ¼ 88) Liraglutide Group (n ¼ 43) Detemir Group (n ¼ 45) P Value

Sex (male/female) 52/36 23/20 29/16 .40 Age (years) 60.4 Æ 12.0 60.1 Æ 12.6 60.6 Æ 11.5 .99 Body weight (kg) 63.4 Æ 12.3 62.0 Æ 12.4 65.4 Æ 12.2 .24 Body mass index (kg/m2) 24.4 Æ 3.5 24.2 Æ 3.7 24.7 Æ 3.4 .46 Duration of diabetes (years) 9.0 Æ 8.3 9.6 Æ 9.5 8.5 Æ 7.1 .53

HbA1c, % (mmol/mol) 6.5 Æ 0.7 (47.7 Æ 7.6) 6.5 Æ 0.7 (47.9 Æ 7.9) 6.5 Æ 0.7 (47.5 Æ 7.3) .86 Fasting plasma glucose (mg/dL) 131.3 Æ 28.4 133.7 Æ 31.7 128.6 Æ 25.0 .34 Fasting C-peptide (ng/mL) 1.58 Æ 0.90 1.64 Æ 0.97 1.54 Æ 0.76 .67 Total insulin (units) 22.6 Æ 5.7 23.1 Æ 6.5 22.1 Æ 4.9 .40

Parameters described as mean Æ standard deviation at baseline in each group were analyzed using an unpaired t test.

treatments, with 5% significance. The study was not study end point, HbA1c values were 6.7% Æ 1.3% powered for a noninferiority analysis. Likely, the (49.3 Æ 13.6 mmol/mol), Table 2, and 6.8% Æ 1.1% response rates were based on results from previous (50.4 Æ 12.5 mmol/mol) in the liraglutide and detemir clinical trials.17 Assuming a dropout rate of 10%, the final groups, respectively (Table 3). There were no significant sample size was 90 patients. The last-observation-carried- differences in the mean changes of HbA1c between the 2 forward method was used for missing data for the intent- groups (0.1% Æ 0.9% (1.4 Æ 9.8 mmol/mol) in the liraglu- to-treat (ITT) analysis; the end point was defined as the tide group versus 0.3% Æ 0.8% (2.8 Æ 8.8 mmol/mol) in last study visit (from baseline to week 24 visit). the detemir group (P ¼ .46; Table 4). Results were similar The baseline characteristics are summarized using for the ITT population and the 24-week completer means Æ standard deviations (SDs) for continuous vari- population. ables and frequencies for categorical variables. Continu- At 24 weeks, the mean body weight in the liraglutide ous variables were compared using the unpaired t test or and detemir groups changed from 62.0 Æ 12.4 to Mann–Whitney U test and categorical variables using the 60.9 Æ 13.5 kg and from 65.4 Æ 12.2 to 65.8 Æ 13.2 kg, chi-square test or Fisher’s exact test. The unpaired t test respectively (Tables 2 and 3). The respective mean was used to examine the difference in the mean change changes in body weight from the baseline values in these from the baseline to the end point between the 2 groups. groups were À 1.1 Æ 3.2 kg versus 0.4 Æ 2.7 kg (P ¼ .020; The changes from the baseline to the end point in each Table 4). There were no significant differences in the group were analyzed using repeated-measures analysis of mean changes of FPG, fasting, or postprandial SMBG variance. A 2-sided P < .05 was considered statistically from the baseline values, between the 2 groups (Table 4). significant. Statistical analyses were conducted with JMP Patient-reported outcome measures were assessed in 10 (SAS Institute Inc., Cary, North Carolina). 88 patients (43 and 45 patients in the liraglutide and detemir groups, respectively). The overall satisfaction score for the DTSQs in the liraglutide and detemir groups Results changed from 25.2 Æ 7.4 to 29.9 Æ 5.3 (P < .001) and In total, 110 patients were screened, and 90 were 26.4 Æ 6.1 to 28.3 Æ 6.4 (P ¼ .12), respectively (Table 5). randomized to treatment; 88 were treated (ITT popula- No significant differences were observed in the changes in tion), and 82 completed the study. Eighty-eight patients the overall satisfaction score between the 2 groups were included in our analysis. Baseline characteristics (4.6 Æ 1.2 in the liraglutide group versus 1.9 Æ 1.2 in the were similar between the 2 groups (Table 1). Liraglutide detemir group; P ¼ .10). The Hypoglycemia score treatment was discontinued in 2 patients because of improved from 2.1 Æ 1.9 to 1.0 Æ 1.5 (P < .001) and sustained hyperglycemia and in another 2 patients from 2.0 Æ 1.7 to 1.0 Æ 1.3 (P ¼ .004) in the liraglutide because of adverse events (1 loss of appetite, 1 nausea). and detemir groups (Table 5), respectively, but the Detemir treatment was discontinued in 1 patient because differences were not significant. of sustained hyperglycemia and in another patient because Hypoglycemia (2 patients, 4.4% vs 8 patients, 17.8%, of an adverse event (1 foot gangrene; Figure 1). Whereas P ¼ .04) and the rate of hypoglycemic episodes (0.1 events/ 29% of the detemir-treated patients received a titrating patient per year vs 1.9 events/patient per year; P ¼ .02) and dose of detemir (mean, þ3.2 units per day), no liraglutide- gastrointestinal symptoms (19 patients, 42.2% vs 3 treated patients received an addition of glimepiride. patients, 6.7%, P < .001) were observed during the 24 Baseline HbA1c was 6.5% Æ 0.7% (47.9 Æ 7.9 mmol/ weeks in the liraglutide and detemir groups, respectively. mol) and 6.5% Æ 0.7% (47.5 Æ 7.3 mmol/mol) in the Cases of severe hypoglycemia or diabetic ketoacidosis liraglutide and detemir groups, respectively (Table 1). At were not observed in either group during the trial. 4 The Journal of Clinical Pharmacology / Vol XX No XX 2015

110 patients screened 20 patients excluded not eligible for inclusion criteria

90 patients Randomized

Liraglutide group Detemir group 45 patients 45 patients

6 patients excluded 2 patients excluded Lost to follow up 2 Adverse event 1 Adverse event 2 Loss of glucose control 1 Loss of glucose control 2

39 patients completed 43 patients completed 43 Analysed ITT 45 Analysed ITT Figure 1. Flow chart of study participants throughout the trial.

Discussion is more effective in achieving glycemic control than basal In this 24-week open-label, randomized, controlled trial insulin glargine,17 probably because liraglutide improves we assessed the effectiveness of switching from a basal- both postprandial glucose and fasting glucose levels.20 In bolus insulin regimen to once-daily injection with contrast, patients in our study had good or moderate liraglutide versus insulin detemir plus sitagliptin in glycemic control. Under such conditions, improvement in Japanese patients with type 2 diabetes with good or postprandial hyperglycemia seemed to be important for moderate control. Surprisingly, we did not find a improving the HbA1c levels, given that several studies significant difference in glycemic control between treat- attributed a slightly elevated HbA1c level to postprandial ments. We recommend this switch of treatment for hyperglycemia.21,22 Based on the results of previous subjects having good or moderate glycemic control with a studies, liraglutide was expected to reduce fasting and combination injective therapy (refer to Figure 7.1 in the postprandial glucose levels.20,23 In addition, we added ADA Standards of Medical Care in Diabetes 2015). sitagliptin to the detemir group to improve postprandial Many studies have compared the GLP-1 receptor glucose levels in this study based on previous reports of agonist and basal insulin therapy in patients with poorly the efficacy and safety of insulin therapy plus sitaglip- controlled type 2 diabetes17–19 and found that liraglutide tin,14,24 mainly by improving postprandial glucose levels.

Table 2. Time Courses for Clinical Parameters in the Liraglutide Group

0 Weeks 8 Weeks 16 Weeks 24 Weeks P Value

Body weight (kg) 62.0 Æ 12.4 61.1 Æ 12.6 60.9 Æ 13.1 60.9 Æ 13.5 .026a

HbA1c (%) (mmol/mol) 6.5 Æ 0.7 (47.9 Æ 7.9) 6.4 Æ 0.9 (46.2 Æ 9.5) 6.5 Æ 1.1 (47.2 Æ 11.8) 6.7 Æ 1.3 (49.3 Æ 13.6) .15 Fasting plasma glucose (mg/dL) 133.7 Æ 31.7 136.9 Æ 45.3 140.6 Æ 44.0 137.5 Æ 48.5 .71 Self-monitored blood glucose, fasting (mg/dL) 128.7 Æ 28.8 123.3 Æ 25.4 122.7 Æ 28.3 131.4 Æ 44.8 .34 Self-monitored blood glucose, postprandial (mg/dL) 153.8 Æ 26.3 140.1 Æ 28.6 148.5 Æ 32.9 155.8 Æ 49.9 .17

Parameters are described as mean Æ standard deviation at 0, 8, 16, and 24 weeks. The differences in parameters 8, 16, and 24 weeks from baseline (0 weeks) were analyzed using a repeated-measures analysis of variance. aP < .05. Inoue et al 5

Table 3. Time Course for Clinical Parameters in the Detemir Group

0 Weeks 8 Weeks 16 Weeks 24 Weeks P Value

Body weight (kg) 65.4 Æ 12.2 64.8 Æ 12.7 65.6 Æ 13.0 65.8 Æ 13.2 .31 a HbA1c, % (mmol/mol) 6.5 Æ 0.7 (47.5 Æ 7.3) 6.5 Æ 0.8 (47.7 Æ 8.8) 6.7 Æ 1.1 (49.3 Æ 11.8) 6.8 Æ 1.1 (50.4 Æ 12.5) .041 Fasting plasma glucose (mg/dL) 128.6 Æ 25.0 129.1 Æ 32.1 132.8 Æ 38.4 131.6 Æ 42.6 .62 Self-monitored blood glucose, fasting (mg/dL) 126.8 Æ 37.1 124.6 Æ 38.9 123.4 Æ 36.0 118.5 Æ 23.9 .25 Self-monitored blood glucose, postprandial (mg/dL) 134.4 Æ 34.0 139.6 Æ 24.9 142.7 Æ 25.9 140.0 Æ 25.6 .71

Parameters are described as mean Æ standard deviation at 0, 8, 16, and 24 weeks. The differences in parameters at 8, 16, and 24 weeks from baseline (0 weeks) were analyzed using a repeated-measures analysis of variance. aP < .05.

Insulin therapy is associated with weight gain.7 However, hypoglycemia and its potential sequelae. The ratio of treatment with insulin detemir appears to result in less gastrointestinal symptoms was significantly higher in the weight gain than NPH insulin25 or insulin glargine.26 liraglutide group compared with the detemir group. Furthermore, sitagliptin does not appear to have an However, very few patients discontinued the study influence on body weight.27 Conversely, liraglutide was because of such symptoms, which is line with a previous reported to reduce body weight compared with insulin report.17 glargine.17,28 In this study, body weight was significantly A study evaluating the switch from insulin therapy to reduced in the liraglutide group compared with the the GLP-1 receptor agonist exenatide reported that it is detemir group. However, loss of body weight was not feasible to sustain glycemic control when substituting 31 associated with improvement of HbA1c (r ¼ .03, P ¼ .55). insulin with exenatide. In addition, tests to predict the Gastrointestinal symptoms as the side effect were response to switching from insulin to a GLP-1 receptor observed frequently, but its frequency did not differ agonist, such as evaluation of b-cell function using between nonobese patients and overweight/obese patients postprandial C-peptide or the glucagon stimulation test, (P ¼ .48). In addition, loss of body weight did not differ are useful for determining the efficacy of liraglutide between patients with gastrointestinal symptoms and introduction in patients with type 2 diabetes.32,33 In our those without gastrointestinal symptoms (À 0.5 Æ 0.8 kg, study, some patients had difficulty with treatment À 1.6 Æ 0.6 kg, respectively; P ¼ .27). Therefore, we continuation by the strict glycemic control achieved in suggest that gastrointestinal symptom was not associated both treatment groups. We plan to conduct a subanalysis on with weight reduction. In East Asians, even a small potential factors triggering glycemic control aggravations. decrease in insulin secretory function may lead to a rapid As in our study, the introduction of liraglutide decrease in the threshold level of insulin resistance above improved patient satisfaction in DTSQs in previous which type 2 diabetes occurs.29 Comparison of patients studies.34,35 In another study, the improvement in with a body mass index (BMI) greater than 22 (ideal BMI treatment satisfaction was greater for insulin glargine in Japanese) and those with a BMI less than 22 revealed once daily than for insulin lispro 3 times per day.36 Thus, that loss of body weight was indistinguishable between the reduction in number of injection times was directly the 2 groups (À 0.9 Æ 0.6 kg, À 1.9 Æ 1.0 kg, respectively; related to patient satisfaction improvement. In a study of P ¼ .38). exenatide and insulin glargine combined with oral Recently, it was reported that hypoglycemia is medications among patients with type 2 diabetes, associated with cardiovascular events and cognitive significant improvements were observed in patient- function.30 Therefore, switching from a basal-bolus reported outcomes,37 but patient-reported outcomes insulin regimen to liraglutide may reduce the risk of were not reported when they switched from 4 injections

Table 4. Change From Baseline and Difference Between Groups

Liraglutide Detemir Difference Between Between-Group Group Group Groups P Value

Body weight (kg) À 1.1 Æ 3.2 0.4 Æ 2.7 À 1.5 Æ 0.6 .020a

HbA1c, % (mmol/mol) 0.1 Æ 0.9 (1.4 Æ 9.8) 0.3 Æ 0.8 (2.8 Æ 8.8) À 0.1 Æ 0.2 (-1.5 Æ 2.0) .46 Fasting plasma glucose (mg/dL) 3.8 Æ 40.5 3.1 Æ 36.0 0.8 Æ 8.2 .93 Self-monitored blood glucose, fasting (mg/dL) À 7.8 Æ 20.6 À 9.3 Æ 19.7 1.5 Æ 6.1 .80 Self-monitored blood glucose, postprandial (mg/dL) À 8.3 Æ 24.9 3.4 Æ 32.6 À 11.7 Æ 10.1 .26 aChanges in parameters described as mean Æ standard deviation between 0 and 24 weeks were analyzed using an unpaired t test. 6 The Journal of Clinical Pharmacology / Vol XX No XX 2015

Table 5. DTSQs in the 2 Groups

Liraglutide Group Detemir Group

0 Weeks 24 Weeks P Value 0 Weeks 24 Weeks P Value

1. Current treatment 4.7 Æ 1.3 5.2 Æ 1.0 .023a 4.7 Æ 1.2 4.9 Æ 1.4 .53 2. Hyperglycemia 2.2 Æ 1.6 2.0 Æ 1.7 .37 2.3 Æ 1.8 2.4 Æ 1.8 .78 3. Hypoglycemia 2.1 Æ 1.9 1.0 Æ 1.5 <.001c 2.0 Æ 1.7 1.0 Æ 1.3 .004c 4. Convenience 3.7 Æ 1.7 5.1 Æ 1.0 < .001 4.1 Æ 1.6 4.7 Æ 1.4 .04b 5. Flexibility 3.9 Æ 1.6 4.6 Æ 1.3 .016b 4.0 Æ 1.3 4.6 Æ 1.4 .05 6. Understanding 4.6 Æ 1.1 5.0 Æ 1.1 .08 4.4 Æ 1.3 4.4 Æ 1.4 .92 7. Recommend 4.3 Æ 1.8 5.0 Æ 1.1 .008c 4.6 Æ 1.2 4.8 Æ 1.5 .47 8. Continue 4.1 Æ 1.8 5.0 Æ 1.2 .005c 4.5 Æ 1.5 4.9 Æ 1.3 .15 9. Overalla 25.2 Æ 7.4 29.9 Æ 5.3 <.001c 26.4 Æ 6.1 28.3 Æ 6.4 .12

Values are expressed as means Æ standard deviations. The differences in parameters 24 week from baseline (0 weeks) were analyzed using a paired t test. aCombined score of 1, 4, 5, 6, 7, and 8; bP < .05; cP < .01. to once-daily injection. In a Japanese study, the mean of ketoacidosis, loss of glycemic control, and other overall satisfaction scores were 24.2 Æ 8.1, 28.5 Æ 6.6, adverse events. Second, the baseline BMI values in this and 28.4 Æ 7.0 in the insulin, oral hypoglycemic agent, study were lower than those of the Western countries. It and diet therapy groups, respectively.16 In the present has been reported that East Asians have smaller mean study, the mean overall satisfaction scores in both groups height, weight, and BMI and less mean total visceral fat were similar to those of the insulin therapy group and volume than blacks/African Americans and whites.39–42 lower than those of the oral hypoglycemic agent group. In the United States, patients with type 2 diabetes The result of the DTSQ evaluation showed that patient experienced statistically significant decreases in HbA1c satisfaction level with treatment improved significantly and body weight after initiating liraglutide regardless of after switching to liraglutide, although glucose control their BMI,43 but their BMIs were all greater than 25.0. In (HbA1c level) was not different. The achievement of our study, the mean BMI was 24.4 Æ 3.5. Observational HbA1c 7% (53 mmol/mol) seemed unrelated to DTSQ studies performed in Japan have revealed that sitagliptin score improvement. On the other hand, the mean overall is more effective for patients with a lower BMI.44,45 In the satisfaction scores in the detemir group were not present trial, liraglutide and detemir lowered HbA1c improved significantly. In the liraglutide group, all items further in patients with a BMI < 25 kg/m2 than in those without “understanding” in DTSQs were improved from with a BMI > 25 kg/m2, and the results were not the baseline values. The “convenience” score was statistically significant. The difference in the BMIs among significantly improved, even in the detemir group. We the Japanese, American, and European patients was likely considered that improvement of the convenience score in the cause of the different results. Third, all the patients in the 2 groups was caused by reduction in the number of this trial were Japanese, as the maximum allowable injections. As for other factors, we thought the differences liraglutide dose in Japan is 0.9 mg. between the 2 groups may be caused by the frequency of hypoglycemia, loss of body weight, and aggravation of glucose control in the detemir group. Conclusions The effect of switching from a basal-bolus insulin In conclusion, there were no significant differences in therapy to liraglutide or detemir plus sitagliptin was glycemic control between liraglutide and insulin detemir essentially similar in the elderly patients (age older than plus sitagliptin, after switching from a basal-bolus insulin 70; data not shown). However, more elderly patients are regimen in Japanese patients with type 2 diabetes with needed to make this conclusion. good or moderate glycemic control. However, liraglutide Our study has some limitations. First, our patients were reduced body weight significantly and had a lower ratio of limited to using a total insulin dose of 30 units and to hypoglycemia, compared with insulin detemir plus maintain endogenous insulin-secreting capacity before sitagliptin. Moreover, the switch a from basal-bolus switching. Furthermore, our patients had good or insulin regimen to liraglutide improved patient satisfac- moderate glycemic control (HbA1c 7.3% [56 mmol/ tion without loss of glycemic control. Patients with good mol]). Because in a previous study, a patient was or moderate glycemic control could benefit from switch- diagnosed with ketoacidosis after switching to liraglu- ing to a liraglutide once-daily regimen from a basal-bolus tide,38 we selected our patients carefully to avoid the risk insulin regimen. Inoue et al 7

Acknowledgments 08. Action to Control Cardiovascular Risk in Diabetes Study Group, Gerstein HC, Miller ME, Byington RP, et al. Effects of We thank all the physicians who participated in this study. We intensive glucose lowering in type 2 diabetes. N Engl J Med. 2008; thank Hitoshi Ishii from the Department of Diabetology, Nara 358:2545–2559. Medical University, for his valuable help with the DTSQ 09. Gregg EW, Mangione CM, Cauley JA, et al. Diabetes and incidence Japanese version. The authors thank Enago (http://www.enago. of functional disability in older women. Diabetes Care. 2002; jp) for the English-language review. 25:61–67. 10. Ott A, Stolk RP, van Harskamp F, et al. Diabetes mellitus and Declaration of Conflicting Interests the risk of dementia: The Rotterdam Study. Neurology. 1999;53: 1937–1942. Yasuo Terauchi received honoraria for lectures from MSD K.K.; 11. Cukireman T, Gerstein HC, Williamson JD. Cognitive decline and Ono Pharmaceutical Co, Ltd; Nippon Boehringer Ingelheim Co, dementia indiabetes—systematic overview of prospective observa- Ltd; Novartis Pharma K.K.; Takeda Pharmaceutical Co, Ltd; tional studies. Diabetologia. 2005;48:2460–2469. Mitsubishi Tanabe Pharma Corp; Daiichi Sankyo Co, Ltd; 12. Vilsbùll T, Zdravkovic M, Le-Thi T, et al. Liraglutide, a long-acting Sanwa Kagaku Kenkyusho Co, Ltd; Kowa Pharmaceutical Co, human glucagon-like peptide-1 analog, given as monotherapy significantly improves glycemic control and lowers body weight Ltd; Novo Nordisk Pharma Ltd; Eli Lilly Japan K.K.; Sanofi K. without risk of hypoglycemia in patients with type 2 diabetes. K.; Shionogi & Co, Ltd; Bayer Yakuhin, Ltd; and AstraZeneca Diabetes Care. 2007;30:1608–1610. K.K. and obtained research support from MSD K.K.; Ono 13. Vilsboll T, Christensen M, Junker AE, Knop FK, Gluud LL. Effects Pharmaceutical Co, Ltd; Nippon Boehringer Ingelheim Co, Ltd; of glucagon-like peptide-1 receptor agonists on weight loss: Novartis Pharma K.K.; Takeda Pharmaceutical Co, Ltd; systematic review and meta-analyses of randomized controlled trials. BMJ. 2012;344:d7771. Mitsubishi Tanabe Pharma Corp; Daiichi Sankyo Co, Ltd; 14. Takai M, Ishikawa M, Maeda H, et al. Safety and efficacy of adding Sanwa Kagaku Kenkyusho Co, Ltd; Novo Nordisk Pharma Ltd; sitagliptin to insulin in patients with type 2 diabetes: the ASSIST-K Eli Lilly Japan K.K.; Sanofi K.K.; Dainippon Sumitomo Pharma study. Diabetes Res Clin Pract. 2014;103:e30–e33. Co, Ltd; Shionogi & Co, Ltd; Bayer Yakuhin, Ltd; Astellas 15. Bradley C. The diabetes treatment satisfaction questionnaire: Pharma, Inc; Pfizer Japan, Inc; and AstraZeneca K.K. Yuichiro DTSQ. In: Bradley C (ed) Handbook of Psychology and Diabetes: A Guide to Psychological Measurement in Diabetes Research and Inoue, Akinobu Nakamura, Yoshinobu Kondo, Kumiko Practice. Chur, Switzerland: Harwood Academic Publishers; Hamano, and Shinobu Satoh declare that they have no conflicts 1994:111–132. of interest. 16. Ishii H, Bradley C, Riazi A, Barendse S, Yamamoto T. The Japanese version of the Diabetes Treatment Satisfaction Questionnaire References (DTSQ): translation and clinical evaluation. J Clin Exp Med. 01. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of 2000;192:809–814 [in Japanese]. hyperglycemia in type 2 diabetes: a patient-centered approach: 17. Russell-Jones D, Vaag A, Schmitz O, et al. Liraglutide Effect and position statement of the American Diabetes Association (ADA) Action in Diabetes 5(LEAD-5) metþSU Study Group. Liraglutide and the European Association for the Study of Diabetes (EASD). vs insulin glargine and placebo in combination with metformin Diabetologia. 2012;55:1577–1596. and sulfonylurea therapy in type 2 diabetes mellitus (LEAD-5 02. Giugliano D, Maiorino MI, Bellastella G, et al. Efficacy of insulin metþSU): a randomized controlled trial. Diabetologia. 2009; analogs in achieving the hemoglobin A1c target of <7% in type 2 52:2046–2055. diabetes: meta-analysis of randomized controlled trials. Diabetes 18. Davies M, Heller S, Sreenan S, et al. Once-Weekly Exenatide Care. 2011;34:510–517. Versus Once-or Twice-Daily Insulin Detemir: a randomized 03. Shichiri M, Kishikawa H, Ohkubo Y, Wake N. Long-term results of controlled trial. Diabetes Care. 2013;36:1368–1376. the Kumamoto Study on optimal diabetes control in type 2 diabetic 19. Li WX, Gou JF, Tian JH, Yan X, Yang L. Glucagon-like peptide-1 patients. Diabetes Care. 2000;23(Suppl 2): B21– B29. receptor agonists versus insulin glargine for type 2 diabetes 04. Ohkubo Y, Kishikawa H, Arai E, et al. Intensive insulin therapy mellitus: a systematic review and meta-analysis of randomized prevents the progression of diabetic microvascular complications in controlled trials. Curr Ther Res Exp. 2010;71:211–238. Japanese patients with non-insulin-depandent diabetes mellitus: a 20. Flint A, Kapitza C, Hindsberger C, Zdravvkovic M. The once-daily randomized prospective 6-year study. Diabetes Res Clin Pract. human glucagon-like peptide-1 (GLP-1) analog liraglutide im- 1995;28:103–117. proves postprandial glucose levels in type 2 diabetes patients. Adv 05. United Kingdom Prospective Diabetes Study (UKPDS) Group. Ther. 2011;3:213–226. Intensive blood-glucose control with sulphonylureas or insulin 21. Monnier L, Lapinski H, Colette C. Contributions of fasting and compared with conventional treatment and risk of complications in postprandial plasma glucose increments to the overall diurnal patients with type 2 diabetes (UKPDS 33). Lancet. 1998;352:837– hyperglycemia of type 2 diabetic patients: variations with increasing 853. levels of HbA 1c). Diabetes Care. 2003;26:881–885. 06. United Kingdom Prospective Diabetes Study (UKPDS) Group. 22. Kikuchi K, Nezu U, Shirakawa J, et al. Correlations of fasting and United Kingdom Prospective Diabetes Study 24: a 6-year, postprandial blood glucose increments to the overall diurnal randomized, controlled trial comparing sulfonylurea, insulin, and hyperglycemic status in type 2 diabetic patients: variations with metformin therapy in patients with newly diagnosed type 2 diabetes levels of HbA1c. Endocr J. 2010;57:259–266. that could not be controlled with diet therapy. Ann Intern Med. 23. Mori Y, Taniguchi Y, Sezaki k, Yokoyama J, Utsunomiya K. 1998;128:165–175. Liraglutide narrows the range of circadian glycemic variations 07. The Diabetes Control and Complications Trial (DCCT) Research in Japanese type 2 diabetes patients and nearly flattens these Group. The effect of intensive treatment of diabetes on the variations in drug-na€ıve type 2 diabetes patients: a continuous development and progression of long-term complications in insulin- glucose monitoring-based study Diabetes Technol Therapy. dependent diabetes mellitus. N Engl J Med. 1993;329:977–986. 2011;11:1139–1144 8 The Journal of Clinical Pharmacology / Vol XX No XX 2015

24. Vilsbùll T, Rosenstock J, Yki-J€arvinen H, et al. Efficacy and safety diabetes compared with sitagliptin, each as an add on to metformin. of sitagliptin when added to insulin therapy in patients with type 2 Diabet Med. 2011;28:333–337. diabetes. Diabetes Obes Metab. 2010;12:167–177. 36. Bretzel RG, Nuber U, Landgraf W, et al. Once-daily basal insulin 25. Philis-Tsimikas A, Charpentier G, Clauson P, et al. Comparison of glargine versus thrice-daily prandial insulin lispro in people once-daily insulin detemir with NPH insulin added to a regimen of with type 2 diabetes on oral hypoglycaemic agents (APOLLO): oral antidiabetic drugs in poorly controlled type 2 diabetes. Clin an open randomized controlled trial. Lancet. 2008;371:1073– Ther. 2006;28:1569–1581. 1084. 26. Swinneg SG, Simon AC, Holleman F, Hoekstra JB, Devries JH. 37. Secnik BK, Matza LS, Oglesby A, et al. Patient-reported outcomes Insulin detemir versus insulin glargine for type 2 diabetes mellitus. in a trial of exenatide and insulin glargine for the treatment of type 2 Cochrane Database Syst Rev. 2011;7:CD006383. diabetes. Health Qual Life Outcomes. 2006;4:80. 27. Iwamoto Y Taniguchi T, Nonaka K, et al. Dose-ranging efficacy of 38. Gooderick D, Dashora U, Kumar S. Ketoacidosis in type 2 diabetes- sitagliptin, a dipeptidyl peptidase-4 inhibitor, in Japanese patients is it type 1 and 1/2 diabetes? BMJ Case Rep. 2011. doi: 10.1136/ with type 2 diabetes mellitus. Endocr J. 2010;57:383–394. bcr.07.2011.4460 28. Kozawa N, Inoue K, Iwamoto R, et al. Liraglutide is effective in 39. Lovejoy JC, de la Bretonne JA, Klemperer M, Tulley R. Abdominal type 2 diabetic patients with sustained endogenous insulin-secreting fat distribution and metabolic risk factors: effects of race. capacity. J Diabetes Investig. 2011;3:294–297. Metabolism. 1996;45:1119–1124. 29. Kodama K, Tojjar D, Yamada S, et al. Ethnic differences in the 40. Van Horn LV, Ballew C, Liu K, et al. Diet, body size, and plasma relationship between insulin sensitivity and insulin response: a syste- lipids-lipoproteins in young adults: differences by race and sex. The matic review and meta-analysis. Diabetes Care. 2013;36:1789–1796. Coronary Artery Risk Development in Young Adults (CARDIA) 30. Whitmer RA, Karter AJ, Yaffe K, Quesenberry CP Jr, Selby JV. study. Am J Epidemiol. 1991;133:9–23. Hypoglycemic episodes and risk of dementia in older patients with 41. Ortiz O, Russell M, Daley TL, et al. Differences in skeletal muscle type 2 diabetes mellitus. JAMA. 2009;301:1565–1572. and bone mineral mass between black and white females and their 31. Davis SN, Johns D, Maggs D, et al. Exploring the substitution of relevance to estimates of body composition. Am J Clin Nutr. exenatide for insulin in patients with type 2 diabetes treated with 1992;55:8–13. insulin in combination with oral antidiabetes agents. Diabetes Care. 42. Taaffe DR, Cauley JA, Danielson M, et al. Race and sex effects on 2007;30:2767–2772. the association between muscle strength, soft tissue, and bone 32. Iwao T, Sakai K, Sata M. Postprandial serum C-peptide is a useful mineral density in healthy elders: the Health, Aging, and Body parameter in the prediction of successful switching to liraglutide Composition Study. J Bone Miner Res. 2001;16:1343–1352. monotherapy from complex insulin therapy in Japanese patients 43. Chitnis AS, Ganz ML, Benjamin N, Langer J, Hammer M. Clinical with type 2 diabetes. J Diabetes Complications. 2013;27:87–91. effectiveness of liraglutide across body mass index in patients with 33. Kondo Y, Satoh S, Nagakura J, et al. Defining criteria for the type 2 diabetes in the United States: a retrospective cohort study. introduction of liraglutide using the glucagon stimulation test in Adv Ther. 2014;31:986–999. patients with type 2 diabetes. J Diabetes Investig. 2013;4:571–575. 44. Maeda H, Kubota A, Tanaka Y, et al. ASSET-K Study group. The 34. Lind M, Jendle J, Torffvit O, Lager I. Glucagon-like peptide 1 safety, efficacy and predictors for HbA1c reduction of sitagliptin in (GLP-1) analogue combined with insulin reduces HbA1c and the treatment of Japanese type 2 diabetes. Diabetes Res Clin Pract. weight with low risk of hypoglycemia and high treatment 2012;95: e20–e22. satisfaction. Prim Care Diabetes. 2012;6:41–46. 45. Kubota A, Maeda H, Kanamori A, et al. Pleiotropic effects of 35. Davies M, Pratley R, Hammer M, Thomsen AB, Cuddihy R. sitagliptin in the treatment of type 2 diabetes mellitus patients J Clin Liraglutide improves treatment satisfaction in people with Type 2 Med Res. 2012;4(5):309–313