Diabetes Care 1

Tomoya Mita,1 Naoto Katakami,2,3 Sitagliptin Attenuates the Toshihiko Shiraiwa,4 Hidenori Yoshii,5 Tomio Onuma,5 Nobuichi Kuribayashi,6 Progression of Carotid Intima- Takeshi Osonoi,7 Hideaki Kaneto,2 Keisuke Kosugi,8 Yutaka Umayahara,9 Media Thickening in Insulin- Tsunehiko Yamamoto,10 Kazunari Matsumoto,11 Treated Patients With Type 2 Hiroki Yokoyama,12 Mamiko Tsugawa,13 Masahiko Gosho,14 Iichiro Shimomura,2 Diabetes: The Sitagliptin and Hirotaka Watada,1 on behalf of the Collaborators on the Sitagliptin Prevention Preventive Study of Intima-Media Study of Intima-Media Thickness Thickness Evaluation (SPIKE) Evaluation (SPIKE) Trial*

A Randomized Controlled Trial 1Department of Metabolism & Endocrinology, Juntendo University Graduate School of Medi- DOI: 10.2337/dc15-2145 cine, Bunkyo-ku, Tokyo, 2Department of Metabolic Medicine, Uni- versity Graduate School of Medicine, Suita, Osaka, Japan 3Department of Atherosclerosis and Metabo- lism, Osaka University Graduate School of Med- icine, Suita, Osaka, Japan 4Shiraiwa Medical Clinic, Kashiwara, Osaka, Japan OBJECTIVE 5Department of Medicine, Diabetology & Endo- The effect of additional treatment with oral hypoglycemic agents on the progres- crinology, Juntendo Tokyo Koto Geriatric Medi- cal Center, Koto-ku, Tokyo, Japan sion of atherosclerosis remains unknown in insulin-treated patients with type 2 6Misaki Naika Clinic, Funabashi, Chiba, Japan diabetes mellitus (T2DM). We assessed the effects of sitagliptin, a dipeptidyl 7Naka Memorial Clinic, Naka City, Ibaraki, Japan 8 peptidase 4 inhibitor, on carotid intima-media thickness (IMT) in T2DM. Osaka Police Hospital, Tennoji-ku, Osaka, Japan 9Osaka General Medical Center, Sumiyoshi-ku, RESEARCH DESIGN AND METHODS Osaka, Japan 10Kansai Rosai Hospital, Amagasaki-shi, Hyogo, This prospective, randomized, open-label, blinded end point, multicenter, parallel Japan group, comparative study included 282 insulin-treated patients with T2DM free 11Diabetes Center, Sasebo Chuo Hospital, Sa- sebo, Nagasaki, Japan RISK METABOLIC AND CARDIOVASCULAR of a history of apparent cardiovascular diseases who were recruited at 12 clinical 12 n Department of Internal Medicine, Jiyugaoka units and randomly allocated to either the sitagliptin group ( = 142) or the control Medical Clinic, Obihiro, Hokkaido, Japan group (n = 140). The primary outcomes were changes in mean and maximum IMT 13Ikeda Municipal Hospital, Ikeda, Osaka, Japan of the common carotid artery measured by echography at the end of a 104-week 14Department of Clinical Trial and Clinical Epide- treatment period. miology, Faculty of Medicine, University of Tsu- kuba, Tsukuba, Ibaraki, Japan RESULTS Corresponding author: Tomoya Mita, tom-m@ Sitagliptin had a more potent glucose-lowering effect compared with the conven- juntendo.ac.jp. tional treatment (20.5 6 1.0% vs. 20.2 6 0.9%; P = 0.004), without increasing Received 30 September 2015 and accepted 8 December 2015. hypoglycemic episodes or body weight. Changes in the mean and left maximum fi Clinical trial reg. no. UMIN000007396, www IMT, but not right maximum IMT, of the common carotid arteries were signi - .umin.ac.jp/ctr/. cantly greater after sitagliptin treatment compared with conventional treatment This article contains Supplementary Data online (20.029 [SE 0.013] vs. 0.024 [0.013] mm [P =0.005];20.065 [0.027] vs. 0.022 at http://care.diabetesjournals.org/lookup/ [0.026] mm [P =0.021];20.007 [0.031] vs. 0.027 [0.031] mm [P = 0.45], respec- suppl/doi:10.2337/dc15-2145/-/DC1. tively). Over 104 weeks, sitagliptin, but not conventional treatment, significantly T.M. and N.Ka. contributed equally to this study. reduced the mean IMT and left maximum IMT of common carotid arteries relative *Collaborators on the Sitagliptin Preventive to the baseline. Study of Intima-Media Thickness Evaluation (SPIKE) Trial can be found in the Supplementary CONCLUSIONS Data online. Sitagliptin attenuated the progression of carotid IMT in insulin-treated patients © 2016 by the American Diabetes Association. Readers may use this article as long as the work is with T2DM free of apparent cardiovascular disease compared with conventional properly cited, the use is educational and not for treatment. profit, and the work is not altered. Diabetes Care Publish Ahead of Print, published online January 28, 2016 2 Effects of Sitagliptin on Atherosclerosis Diabetes Care

Type 2 diabetes mellitus (T2DM) in- RESEARCH DESIGN AND METHODS of each participating institution, in compli- creases the likelihood of developing car- Study Population ance with the Declaration of Helsinki and diovascular disease (CVD), which is one Japanese patients with T2DM who reg- current legal regulations in Japan. Written of the major causes of mortality and ularly attended the outpatient diabetes informed consent was obtained from all morbidity in these patients (1). Recent clinics at 12 institutions in Japan were participants after a full explanation of studies have cast doubt on the benefits asked to participate in this study, as pre- the study. This study is registered in of strict glycemic control, especially using viously described in detail (16). Patients the University Hospital Medical Informa- insulin, on CVD in patients with estab- with T2DM in whom the target of blood tion Network Clinical Trials Registry lished atherosclerosis or long-standing glucose control specified in the Treat- (UMIN000007396), which is a nonprofit T2DM (2–4) for two main reasons: 1)It ment Guide for Diabetes in 2010 (edited organization in Japan, and meets the re- is assumed that frequent episodes of se- by the Japan Diabetes Society) (17) was quirements of the International Commit- vere hypoglycemia might reduce their not achieved despite insulin therapy, in tee of Medical Journal Editors. fi bene cial effects (5), and 2) weight gain addition to dietary/exercise therapy or may adversely affect the prognosis. concomitant therapeutic drugs for Randomization and Study Therefore, reducing insulin dose by stim- T2DM (other than DPP-4 inhibitors) Intervention Patients were registered at the adminis- ulating endogenous insulin secretion and over a period of $3 months, were in- tration office of this trial via the Inter- increasing insulin sensitivity using oral hy- cluded in the study. Patients who with- net; once enrolled, they were randomly poglycemic agents (OHAs) (6,7) may re- drew from previous treatment with assigned to either the sitagliptin group duce these adverse effects of insulin DPP-4 inhibitor for more than 12 weeks or the control group receiving conven- therapy. However, the effects of treat- were also included; additional inclusion ment with insulin plus other traditional tional treatment consisting of drugs criteria included $30 but ,80 years of OHAs on the progression of atherosclero- other than DPP-4 inhibitors. Randomiza- age (regardless of sex) and signing the sis remain largely unknown. New OHAs tion was performed using a dynamic al- consent form for participation in the with the least risk of undeliverable effects location method based on the number study. The following exclusion criteria and multiple beneficial effects on cardio- of insulin injections, with/without con- also were applied: 1) type 1 or second- vascular (CV) profiles when used with in- comitant pioglitazone, age, and sex. ary diabetes; 2) presence of severe in- sulin are essential for the treatment of Patients in the sitagliptin group were fectious disease, before or after surgery, T2DM. started on sitagliptin (25 mg once daily) or severe trauma; 3) history of myocar- Sitagliptin, a dipeptidyl peptidase 4 in addition to ongoing insulin therapy. dial infarction, angina pectoris, cerebral (DPP-4) inhibitor, increases insulin se- The dose of sulfonylurea was tapered stroke, or cerebral infarction; 4) retinop- cretion and suppresses glucagon release when considered clinically appropriate athy requiring laser photocoagulation by protecting the degradation of incre- to avoid hypoglycemia when starting si- tins without increased risk of hypoglyce- and/or vitrectomy or history of these tagliptin. Initiation of treatment with si- mia and clinically relevant weight gain treatments within 1 year; 5)moderate tagliptin 50 mg once daily was permitted (8,9). Furthermore, the addition of DPP-4 or severe renal dysfunction (serum cre- in patients who were not treated with a . inhibitors to ongoing insulin therapy atinine: males, 1.4 mg/dL; females, sulfonylurea. In patients treated with . was reported to have specific advantages 1.2 mg/dL); 6) severe liver dysfunction sitagliptin25or50mgoncedailyfor $ in reducing the frequency of hypoglyce- (aspartate aminotransferase 100 IU/L); 12 weeks, the dose of sitagliptin was mia and weight gain, in addition to the 7) moderate or severe heart failure (New increased to a maximum of 100 mg York Heart Association stage III sever- expected benefits associated with gly- once daily when HbA1c was $7.0% (8.6 cemic control and limiting insulin dose ity or higher); 8)treatmentwithanin- mmol/mol) (17). The participating phy- (10–12). In addition, evidence from pre- cretin preparation, such as other DPP-4 sicians were allowed to reduce sitaglip- clinical studies suggests that DPP-4 inhib- inhibitors, at the start of the study; 9) tin to 25 or 50 mg/day if treatment with itors could have beneficial effects on treatment with drugs that are not con- 50 or 100 mg/day was considered poorly atherosclerosis in both GLP-1-dependent comitantly administrable with incretin tolerated. The insulin dose could also be and -independent manners (13–15). Thus preparations with regard to the national adjusted, with priority given to achieve the addition of DPP-4 inhibitors to insulin health insurance, such as DPP-4 inhibitors, fasting blood glucose or 2-h postpran- therapy is expected to have beneficial ef- at the start of the study; 10)pregnant, dial blood glucose as recommended in fects on CVD in patients with T2DM. lactating, or possibly pregnant women or the Treatment Guide for Diabetes (17). This study is a multicenter, prospec- those planning to become pregnant during In the control group, either increasing tive, randomized, open-label, blinded the study period; 11) medical history of the dose of current therapy (e.g., insu- end point (PROBE) study designed to de- hypersensitivity to investigational drugs; lin) or adding a sulfonylurea, glinide, or termine the effect of adding sitagliptin and 12) those patients judged as ineli- a-glucosidase inhibitor was allowed, to insulin treatment (sitagliptin group) gible by the clinical investigators. with the goal of achieving the target compared with conventional treatment The subjects were screened consecu- value specified in the Treatment Guide (non-sitagliptin group) on carotid artery tively, and patients who met the above for Diabetes (usually HbA1c ,6.9% [8.44 intima-media thickness (IMT) in CVD-free eligibility criteria were asked to participate. mmol/mol]) (17). The addition of other patients with T2DM. IMT, a marker of pro- All patients who agreed to participate were DPP-4 inhibitors or GLP-1 analogs was gression of atherosclerosis, was evaluated enteredintothestudy.Theprotocolwas banned in the control group. Dose ad- by a simple and noninvasive procedure. approved by the institutional review board justment and addition of metformin and care.diabetesjournals.org Mita and Associates 3

pioglitazone during the study were marked on the Meijer arc) as well as yielded absolute mean differences of banned in both groups. In cases of hy- transverse projections, and the site of 0.02 6 0.01 and 0.01 6 0.01 for mean- poglycemia, the dose of insulin and/or greatest thickness, including plaque le- IMT-CCA and max-IMT-CCA, respec- OHA was titrated. The use of antihyper- sions, was identified along the arterial tively. The intraobserver coefficient of lipidemic and antihypertensive drugs walls. IMT represents the distance be- variation for the measurement was was allowed during the study. tween two parallel echogenic lines cor- 1.1% and 0.7%, respectively. responding to the vascular lumen and Observation Variables and Schedule the adventitia. To avoid interreader var- Sample Size The study period was 104 weeks after iability, all scans were electronically Yokoyama et al. (18) reported previously thepatientswereregistered.Allran- stored and e-mailed to the central office that the mean 6 SD rate of increase in domized participants were followed for (IMT Evaluation Committee, Osaka, Japan) carotid IMT in Japanese patients with di- 104 weeks until the end of the study, to be read, in a random order, by a single abetes was 0.034 6 0.054 mm/year and regardless of adherence to or discontin- experienced reader blinded to the clini- that 1% improvement in HbA1c was asso- uation of study medication for any rea- cal characteristics of the patients using ciated with a 0.02-mm/year improve- son. Clinical outcome, adherence, and automated, digital edge-detection soft- ment in IMT. Given these results, it was adverse events (AEs) were confirmed, ware (IntimaScope; Media Cross, Tokyo, assumed that during a 2-year observation and clinical and biochemical data were Japan) (16). The software system aver- period, registration of at least 232 pa- collected at 0, 26, 52, 78, and 104 weeks ages 60 points of IMT values in the seg- tients was required to obtain 80% power after randomization. ment 2 cm proximal to the dilation of the to detect a difference of 0.04 mm in IMT Primary Outcomes carotid bulb (mean-IMT-CCA). In addi- between the two treatment groups, as- The primary study outcomes were tion, the largest IMTs, including plaque suming an SD of 0.108, 15% dropout, changes in mean IMT of the common lesions in the common carotid arteries and a 0.05 level of significance. On the carotid artery (mean-IMT-CCA) and (max-IMT-CCA), were also measured basis of this calculation, the target num- maximum IMT of the common carotid separately. Reproducibility analysis of ber of enrolled patients was set at 274 for artery (max-IMT-CCA) after the 104- replicate measurements in 20 subjects the 2-year registration period. week treatment period. Investigations werecarriedoutatthestartofthe study, after 52 weeks, and after 104 Table 1—Clinical characteristics of patients in the two groups weeks. Parameters Sitagliptin group Conventional group P value Safety Evaluation Male sex 83 (61) 82 (60) 1.00 AEs were recorded during study as de- Age (years) 63.8 6 9.7 63.6 6 1.0 0.90 scribed in the Supplementary Data online. Current smoking 30 (22) 29 (21) 0.22 Hypertension 75 (55) 86 (63) 0.22 Measurement of Carotid IMT Dyslipidemia 91 (66) 84 (61) 0.45 Ultrasonographic scans of the carotid 6 6 artery were performed by expert sono- Duration of diabetes (years) 17.2 8.5 17.3 8.7 0.94 6 6 graphers who were specifically trained to Insulin injections (time/day) 2.9 1.2 2.9 1.2 0.88 perform the prescribed study examina- Use of oral glucose-lowering agents Metformin 49 (36) 48 (35) 1.00 tion, as reported previously (16). To avoid Sulfonylurea 17 (12) 15 (11) 0.85 intersonographer variability, each par- Glinides 2 (1) 19 (14) ,0.001 ticipant was examined by the same Thiazolidinediones 13 (9) 11 (8) 0.83 sonographer with the same equipment a-Glucosidase inhibitor 41 (30) 42 (31) 1.00 (high-resolution B-mode ultrasound Use of antihypertensive drugs scanner equipped with a high-frequency ACE inhibitors 8 (6) 4 (3) 0.59 [.7.5-MHz] linear transducer, with a Angiotensin II receptor blockers 53 (39) 69 (50) 0.07 limit of detection of ,0.1 mm) through- Direct renin inhibitor 1 (1) 1 (1) 1.00 Calcium channel blocker 39 (28) 45 (33) 0.51 out all the visits. The mean intrain- fi Diuretic drugs 12 (9) 13 (9) 1.00 vestigator coef cient of variation for a-Adrenergic receptor antagonist 2 (1) 2 (1) 1.00 measurement of mean-IMT-CCA and b-Adrenergic receptor antagonist 1 (1) 2 (1) 1.00 max-IMT-CCA (6 SD) in each institution Others 1 (1) 2 (1) 1.00 was 1.6 6 1.2% and 2.4 6 2.3%, respec- Use of lipid-lowering agents tively. The extracranial common carotid Statins 66 (48) 63 (46) 0.81 artery, the carotid bulb, and the internal Ezetimibe 5 (4) 9 (7) 0.41 carotid artery in the neck were scanned Resins 0 (0) 0 (0) d Fibrates 4 (3) 3 (2) 1.00 bilaterally in at least three different lon- gitudinal projections (anterior, lateral, Use of antithrombotic agents Antiplatelet agents 29 (21) 30 (22) 1.00 and posterior, which approximately cor- Anticoagulants 5 (4) 1 (1) 0.21 responded to 60, 90, 150 degrees for the Others 0 (0) 0 (0) d right carotid artery, and 210, 270, and Data are n (%) of patients or mean 6 SD values, unless otherwise indicated. 300 degrees for the left carotid artery 4 Effects of Sitagliptin on Atherosclerosis Diabetes Care

Statistical Analysis groups using Fisher exact test. All sta- mean-IMT-CCA and left max-IMT-CCA, Efficacy was analyzed using mainly the tistical tests were two-sided with a 5% but not right max-IMT-CCA, at 104 weeks full analysis set based on the intent-to- significance level. All analyses were per- (i.e., primary end points of the study) treat principle and secondarily using the formed using SAS software version 9.4 compared with conventional treatment protocol set. Primary analysis was per- (SAS Institute, Cary, NC). (Table 2). Similar findings were noted formed using a mixed-effects model for even in the adjusted mixed-effect mod- repeated measures, with treatment group, RESULTS els, including model 1 (age and sex); time (weeks), interactions between treat- Participants model 2 (model 1 plus BMI, HbA1c,to- ment group and time (weeks), and baseline A total of 282 participants were ran- tal cholesterol, HDL cholesterol, triglyc- IMT as fixed effects; an unstructured co- domly allocated to either the sitagliptin erides, and systolic blood pressure); variate was used to model the covari- group (n = 142) or the conventional ance of within-subject variability. For treatment group (n = 140). After exclud- model 3 (model 2 plus smoking, ACE/ the occurrence of CV events, time to on- ing 8 patients from analyses (they with- angiotensin II receptor blocker, statin, set (one of the secondary end points) drew from the study and/or objected to and antiplatelets); and model 4 (model was analyzed using a log-rank test and the inclusion of their data in the analy- 3 plus OHA) (Supplementary Table 1). Cox proportional hazards model. sis), 137 in the sitagliptin group and 137 Also, ANCOVA models that included Baseline and follow-up group com- in the conventional treatment group treatment group, age, sex, baseline parisons were assessed with the Stu- were included in the full analysis set IMT, systolic blood pressure, and ad- dent t test or Wilcoxon rank sum test (Supplementary Fig. 1). As shown in Ta- ministration of statins produced findings for continuous variables, and the Fisher bles 1, 2, and 3, baseline clinical charac- similar to those of the mixed-effects exact test for categorical variables. teristics, including potential risk factors models (Supplementary Table 2). Over Changes from baseline to treatment vis- for carotid atherosclerosis, were compa- its were assessed with the one-sample 104 weeks, sitagliptin treatment, but rable between the groups. fi t test and Wilcoxon signed rank test not conventional treatment, signi - within the group. The number and per- IMT of the Common Carotid Artery cantly reduced the mean-IMT-CCA and centage of patients reporting AEs were In the mixed-effects model for repeated left max-IMT-CCA, but not right max- presented by treatment group and com- measures in the full-set analysis, sitagliptin IMT-CCA, relative to baseline values pared between the two treatment significantly slowed the worsening of the (Table 2).

Table 2—Effects of sitagliptin on IMT Treatment effect (sitagliptin-conventional treatment) P value Sitagliptin group Conventional group Mean change (95% CI) P value between groups Common mean IMT (mm) Baseline 0.84 6 0.19 (n = 136) 0.84 6 0.21 (n =137) 0.81 52 weeks 0.88 6 0.19 (n = 127) 0.87 6 0.24 (n =124) 0.82 104 weeks 0.81 6 0.19 (n = 121) 0.86 6 0.24 (n =122) 0.07 Mean change (SE) 52 weeks 0.027 (0.013)* 0.028 (0.014)* 20.001 (20.039 to 0.036) 0.94 104 weeks 20.029 (0.013)* 0.024 (0.013) 20.053 (20.090 to 20.016) 0.005 Right maximum IMT (mm) Baseline 1.04 6 0.29 (n = 136) 1.06 6 0.40 (n =137) 0.69 52 weeks 1.09 6 0.42 (n = 127) 1.08 6 0.42 (n =123) 0.78 104 weeks 1.03 6 0.34 (n = 121) 1.09 6 0.49 (n =122) 0.29 Mean change (SE) 52 weeks 0.049 (0.031) 0.013 (0.032) 0.036 (20.052 to 0.123) 0.42 104 weeks 20.007 (0.031) 0.027 (0.031) 20.033 (20.121 to 0.054) 0.45 Left maximum IMT (mm) Baseline 1.10 6 0.32 (n = 137) 1.11 6 0.41 (n =137) 0.87 52 weeks 1.15 6 0.42 (n = 127) 1.12 6 0.45 (n =124) 0.67 104 weeks 1.03 6 0.30 (n = 121) 1.12 6 0.42 (n =122) 0.07 Mean change (SE) 52 weeks 0.038 (0.026) 0.021 (0.026) 0.016 (20.057 to 0.089) 0.66 104 weeks 20.065 (0.027)* 0.022 (0.026) 20.087 (20.161 to 20.014) 0.021 Data are mean 6 SD or mean change (SE) unless otherwise stated. Comparisons of IMTs during treatment with those at baseline were performed using a one-sample t test based on a mixed-effects model for repeated measures. Differences in IMT between groups at each point were analyzed using the Student t test. Differences in change in IMT from baseline at 52 and 104 weeks between groups at each point (treatment effect) were analyzed with a mixed-effects model for repeated measures. Treatment group, week, interactions between treatment group and week, and baseline IMT were included as fixed effects. *P , 0.05. care.diabetesjournals.org Mita and Associates 5

Secondary Outcomes Table 3—Effects of sitagliptin on BMI, glucose metabolism, lipid metabolism, and blood pressure Although both types of treatment signif- Parameters Sitagliptin group Conventional group P value icantly reduced HbA1c levels from base- 2 line, the improvement in HbA1c (value at BMI (kg/m ) 2 6 6 end of study value at baseline) was At baseline 25.0 4.3 (n = 137) 25.0 3.8 (n = 137) 0.88 fi Change from signi cantly greater in the sitagliptin baseline group (20.5 6 1.0%) than the conven- 26 weeks 0.1 6 1.4 (n = 135) 0.1 6 0.9 (n = 136) 0.66 tional group (20.2 6 0.9%; P = 0.004) 52 weeks 0.2 6 1.1 (n = 128) 0.1 6 1.2 (n = 127) 0.75 (Table 3). Similarly, there was a signifi- 78 weeks 0.2 6 1.3 (n = 121)* 0.1 6 1.1 (n = 122) 0.23 cant difference in the change in serum 6 6 104 weeks 0.1 1.5 (n = 119) 0.1 1.8 (n = 117) 0.81 C-peptide concentrations (value at end HbA1c of study 2 value at baseline); serum 6 6 At baseline (%) 8.1 1.1 (n = 137) 8.0 1.0 (n = 137) 0.45 C-peptide concentration was significantly At baseline (mmol/mol) 64.9 6 11.9 (n = 137) 63.9 6 10.6 (n = 137) 0.45 increased in the sitagliptin group but not Change from in the conventional group (Table 3). baseline There were no differences in the value (HbA1c %) and changes in fasting blood glucose 26 weeks 20.5 6 0.9 (n =136)§ 20.3 6 1.0 (n = 136)# 0.02 concentrations at each visit during the 2 6 2 6 52 weeks 0.5 1.0 (n =130)§ 0.4 1.0 (n = 129)§ 0.28 observation period. Notably, the use of 78 weeks 20.5 6 1.0 (n =123)§ 20.3 6 0.9 (n = 126)# 0.09 glinides was significantly higher during 104 weeks 20.5 6 1.0 (n =121)§ 20.2 6 0.9 (n = 122)* 0.004 the study, and the use of a-glucosidase Fasting blood glucose (mmol/L) inhibitors tended to be higher among At baseline 8.64 6 2.85 (n = 136) 8.42 6 2.55 (n = 137) 0.51 the conventional group than among Change from the sitagliptin group at the end of the baseline study (Supplementary Table 3). 26 weeks 20.21 6 3.19 (n = 134) 20.18 6 2.80 (n = 134) 0.95 Similarly, there were no differences 2 6 2 6 52 weeks 0.67 3.54 (n =129)* 0.26 2.36 (n = 129) 0.28 between the two groups in other risk 78 weeks 20.54 6 3.69 (n = 122) 0.00 6 3.12 ( n = 126) 0.21 104 weeks 20.50 6 3.35 (n = 120) 20.40 6 2.64 (n = 122) 0.79 factors for atherosclerosis, such as BMI, blood pressure, and lipid parame- C-peptide (ng/mL) At baseline 1.2 6 0.8 (n = 137) 1.2 6 0.7 (n = 136) 0.61 ters, at baseline or in their changes dur- Change from ing the observation period (Table 3). baseline Blood pressure and lipid metabolism 52 weeks 0.1 6 0.8 (n =126) 0.16 0.7 (n = 123)* 0.82 were well controlled in both groups 104 weeks 0.3 6 0.8 (n = 122)§ 0.0 6 0.6 (n = 121) 0.02 throughout the study. There were also Total cholesterol no significant differences in the use (mmol/L) of antihypertensive drugs and lipid- At baseline 5.02 6 0.91 (n = 136) 4.94 6 0.86 (n = 137) 0.50 lowering agents (Supplementary Table Change from fi baseline (%) 4). Also, no signi cant changes were ob- 26 weeks 22.0 6 14.1 (n =131) 21.0 6 14.3 (n = 127) 0.57 served in the levels of various markers of 52 weeks 22.7 6 13.4 (n =129)* 20.9 6 12.4 (n = 129) 0.26 inflammation and endothelial damage 78 weeks 21.2 6 14.9 (n =122) 0.36 13.3 (n = 123) 0.43 (Table 4). 104 weeks 22.7 6 15.6 (n =121) 21.8 6 14.6 (n = 122) 0.63 During the study, 123 patients (65 in LDL cholesterol the sitagliptin group and 58 in the con- (mmol/L) ventional group) developed AEs, and 17 At baseline 2.85 6 0.78 (n = 137) 2.78 6 0.70 (n = 136) 0.44 patients (8 and 9, respectively) devel- Change from baseline (%) oped serious AEs. There were no signif- 52 weeks 1.4 6 21.4 (n =126) 4.96 21.6 (n = 121)* 0.21 icant differences in the incidence of any 104 weeks 0.9 6 23.4 (n =122) 5.26 23.9 (n = 121)* 0.16 AEs and serious AEs between the two HDL cholesterol groups. We recorded 104 hypoglycemic (mmol/L) events (in 52 patients in the sitagliptin At baseline 1.46 6 0.37 (n = 136) 1.39 6 0.38 (n = 137) 0.14 group and 52 patients in the conventional Change from treatment group) (Supplementary Table baseline (%) fi 26 weeks 1.8 6 13.5 (n =136) 4.36 16.9 (n = 134) 0.18 5). There was no signi cant difference in 52 weeks 22.4 6 13.4 (n = 130)* 0.2 6 15.4 (n = 128) 0.14 the average number of hypoglycemic 78 weeks 2.9 6 15.3 (n = 123)* 5.5 6 16.3 (n = 125)§ 0.21 events between the two groups (0.34 6 104 weeks 0.2 6 15.2 (n =121) 20.5 6 14.7 (n = 121) 0.74 0.85 times/month/person in the sitagliptin 6 Continued on p. 6 group vs. 0.36 0.80 times/month/per- son in the conventional treatment group). One severe hypoglycemic event occurred in each group. Five patients 6 Effects of Sitagliptin on Atherosclerosis Diabetes Care

Table 3—Continued patients with T2DM than in others. In Parameters Sitagliptin group Conventional group P value this regard, the higher rate of increase in carotid IMT among insulin-treated Triglyceride compared with non-insulin-treated pa- (mmol/L) At baseline 1.13 (0.83–1.55) (n =136) 1.17(090–1.72) (n = 137) 0.22 tients with T2DM is probably related to Change from the longer duration of T2DM, poor gly- baseline (%) cemic control, higher frequency of hy- 26 weeks 8.6 (223.4 to 49.0) (n =134)§ 24.3 (224.2 to 25.3) (n = 133) 0.07 poglycemic episodes, increased chance 52 weeks 0.7 (230.0 to 34.1) (n =129) 21.9 (221.0 to 29.8) (n = 129) 0.69 of weight gain, and/or more advanced 78 weeks 26.5 (225.0 to 23.8) (n =122) 26.7 (227.4 to 33.3) (n = 125) 0.94 atherosclerosis. The higher rate of wors- 2 2 2 104 weeks 0.0 ( 25.1 to 44.6) (n =120) 1.6 ( 24.6 to 16.7) (n = 122) 0.35 ening of carotid IMT seems to correlate Systolic blood with the less-than-ideal effect of sita- pressure (mmHg) gliptin on carotid IMT progression at At baseline 130 6 16 (n = 137) 132 6 14 (n = 137) 0.88 52 weeks. In turn, treatment with sita- Change from gliptinresultedinregressionofthe baseline mean-IMT-CCA and left max-IMT-CCA 26 weeks 2 6 16 (n =131) 36 16 (n = 120) 0.55 at the end of the sitagliptin treatment 52 weeks 21 6 17 (n =129) 16 16 (n = 126) 0.39 protocol. Thus long-term treatment 6 6 78 weeks 2 15 (n =123) 3 17 (n = 125)* 0.61 may be required to lessen the worsening 104 weeks 0 6 19 (n =119) 36 17 (n = 122) 0.20 of carotid IMT in insulin-treated patients Diastolic blood pressure with T2DM. (mmHg) In this study, the addition of sitaglip- At baseline 75 6 11 (n =137) 756 12 (n = 137) 0.79 tin to insulin therapy had a sustained Change from glucose-lowering effect amounting to baseline about 0.5% over 2 years from baseline, 26 weeks 22 6 10 (n = 131)* 0 6 12 (n = 120) 0.11 which is almost similar to previous re- 2 6 2 6 52 weeks 1 11 (n =129) 1 12 (n = 126) 0.58 ports (10–12,24). On the other hand, 78 weeks 21 6 12 (n =123) 06 14 (n = 125) 0.51 104 weeks 0 6 11 (n =119) 26 13 (n = 122) 0.16 conventional treatment reduced HbA1c level by only 0.2%, despite our effort to 6 Data are mean SD or median (range) values. Differences in parameters between groups at achieve an optimal glycemic target by baseline were analyzed by the Student t test or Wilcoxon rank sum test. Differences in parameters from baseline to 52 and 104 weeks within the groups were analyzed using a one- adjusting the OHA dosage. In this re- sample t test or Wilcoxon signed rank test. Differences in parameters from baseline to 52 and gard, the dose of insulin was not in- 104 weeks between groups were analyzed by the Student t test or Wilcoxon rank sum test. creased during the study because of §P , 0.001. #P , 0.01. *P , 0.05. concerns about the potential AEs of in- sulin, such as hypoglycemia and weight developed CV events, and 6 patients side than the right side since the left gain. Accordingly, sitagliptin treatment were diagnosed with cancer. There wall of the common carotid artery is was superior to conventional treatment was no significant difference in the in- thicker and more elastic than the right in terms of HbA1c reduction without in- cidence of CVD and cancer between the (19). Therefore, the positive results for creasing the incidence of hypoglycemic two groups. the mean-IMT-CCA and left max-IMT- episodes and weight gain in insulin-treated CCA are more reliable than the negative patients, consistent with previous studies CONCLUSIONS result on the right max-IMT-CCA, which (10–12). However, the difference in HbA1c This is the first PROBE trial to demon- could be a result of the underpowered between the two treatment groups does strate that the addition of sitagliptin to sample. Accordingly, it is feasible to con- not seem to correlate with the reduction insulin treatment significantly dimin- clude with confidence that sitagliptin in carotid IMT, since changes in HbA1c did ished the progression of mean-IMT- has a beneficial effect on carotid IMT not correlate with changes in mean-IMT- CCA and left max-IMT-CCA, but not right in this population. It should be noted CCA and right and left max-IMT-CCA in max-IMT-CCA, compared with conven- that mean-IMT-CCA and left max-IMT- both groups (data not shown). In addition, tional treatment. Because of these dis- CCA were significantly regressed in the sitagliptin treatment still attenuated the crepant results, it might be viewed as an sitagliptin group during the 104-week progression of mean-IMT-CCA and left exaggeration if we were to conclude observation period. max-IMT-CCA compared with conventional that sitagliptin attenuates the progres- Previous studies demonstrated that treatment, even after adjusting for changes sion of carotid IMT. However, the mean- thiazolidinedione drugs had only partial in HbA1c from baseline (Supplementary IMT-CCA is the most reliable parameter effects on the prevention of carotid IMT Table 2). Thus the differences in carotid among these three prespecified primary progression in insulin-treated patients IMT progression could not be explained outcomes of this study: It is generally with T2DM (20,21), whereas they re- by the difference in HbA1c, suggesting believed that the mean-IMT-CCA is duced carotid IMT in non-insulin-treated that these changes in the carotid artery more reproducible compared with patients with T2DM (22,23). Thus it could be independent of the glucose- max-IMT-CCA and that max-IMT-CCA is maybemoredifficult to slow carotid lowering effects. In addition, it is possible more accurately measured on the left IMT progression in insulin-treated that the frequent use of glinides care.diabetesjournals.org Mita and Associates 7

Table 4—Effects of sitagliptin on markers of renal function, inflammation, and and endothelial injury at multiple time endothelial injury points and the progression of carotid Parameters Sitagliptin group Conventional group P value IMT remain unclear in patients with type 1 diabetes (29). Interestingly, a re- eGFR (mL/min/1.73 m2) At baseline 77.7 6 21.2 (n = 137) 79.7 6 24.2 (n =137) 0.47 cent report suggested that evaluation of Change from local inflammation and oxidative stress in baseline the vascular wall, not in serum, was helpful 52 weeks 22.9 6 9.7 (n = 130)§ 22.6 6 10.2 (n = 129)# 0.81 in the full assessment of “non-glycaemic- 104 weeks 24.0 6 10.3 (n =120)§ 22.7 6 11.8 (n = 122)* 0.36 dependent” antiatherosclerotic effects UAE (mg/g creatinine) of DPP-4 inhibitors (30). At baseline 20.3 (8.4–82.2) (n = 136) 19.0 (8.1–83.6) (n =134) 0.64 Although the exact mechanism by Change from which DPP-4 inhibitors induce regression baseline 52 weeks 20.5 (212.7 to 9.1) (n =118) 20.2 (28.7 to 9.2) (n =121) 0.59 of carotid IMT remains uncertain at pre- 104 weeks 0.7 (29.4 to 17.2) (n =114) 20.1 (–11.3 to 8.7) (n =114) 0.37 sent, previous studies postulated a few hs-CRP (ng/dL) scenarios. Inhibition of macrophage accu- At baseline 506 (210–1,310) (n = 137) 487 (277–1,004) (n =136) 0.75 mulation and inflammation by enhancing Change from GLP-1 signaling with sitagliptin or GLP-1 baseline receptor agonists reduced atherosclerotic 52 weeks 217 (2154 to 216) (n =126) 5(2189 to 298) (n = 123) 0.70 lesion formation in a mouse model of ath- 2 2 2 104 weeks 15 ( 142 to 196) (n = 122) 9( 198 to 206) (n =121) 0.71 erosclerosis (13,31). Our group also dem- IL-6 (ng/dL) onstrated that another DPP-4 inhibitor At baseline 1.7 (1.2–2.6) (n = 137) 2.0 (1.2–3.0) (n = 136) 0.34 suppressed macrophage-related inflam- Change from baseline mation and DPP-4-induced smooth mus- 52 weeks 20.1 (20.5 to 0.5) (n =126) 0.0(20.5 to 0.5) (n =123) 0.63 cle proliferation in vitro (14). Furthermore, 104 weeks 0.6 (20.2 to 1.2) (n =122)§ 0.4(20.2 to 1.0) (n = 121)§ 0.53 Nagashima et al. (32) demonstrated that ICAM-1 (ng/mL) GLP-1 suppressed foam cell formation At baseline 224 (180–267) (n = 137) 234 (196–270) (n =136) 0.40 from murine macrophages. These mecha- Change from nisms could potentially contribute to re- baseline duced atherosclerosis. In the future, 52 weeks 22(221 to 19) (n =126) 26(229 to 13) (n =123) 0.06 measurement of local inflammation and 104 weeks 25(223 to 15) (n =122) 27(234 to 21) (n =121) 0.45 the ability to form foam cells in human VCAM-1 (ng/mL) At baseline 728 (636–912) (n = 137) 784 (651–971) (n =160) 0.25 monocytes isolated from peripheral blood Change from may be helpful in investigating the mech- baseline anism behind the beneficial effects of si- 52 weeks 258 (2101 to 32) (n =126)§ 257 (2113 to 20) (n = 123)§ 0.53 tagliptinoncarotidIMT. 104 weeks 212 (275 to 73) (n = 122) 1 (284 to 107) (n =121) 0.50 Three recent randomized clinical stud- Data are mean 6 SD or median (range). Differences in parameters between groups at ies showed that the use of DPP-4 inhibitors baseline were analyzed by the Student t test or Wilcoxon rank sum test. Differences in did not increase nor decrease the CVD parameters from baseline to 52 and 104 weeks within groups were analyzed by the one-sample event rate among patients with T2DM t test or Wilcoxon signed rank test. Differences in parameters from baseline to 52 and 104 weeks between groups were analyzed by the Student t test or Wilcoxon rank sum test. eGFR, compared with placebo (33–35). Since estimated glomerular filtration rate; ICAM-1, intercellular adhesion molecule 1; VCAM-1, these safety outcome studies were origi- vascular cell adhesion molecule 1; IL-6, interleukin 6; UAE, urinary albumin excretion. nally designed as noninferiority clinical tri- *P , 0.05. #P , 0.01. §P , 0.001. als, they probably lacked sufficient power to evaluate the beneficial effects of DPP-4 and a-glucosidase inhibitors in the patients with T2DM (27,28). However, in inhibitors on CVD. In general, the number conventional treatment group over time this study of longer duration, no such of CV events required in a superiority trial affected the results. In this regard, the beneficial effects of sitagliptin on these is larger than that in a noninferiority trial effect of sitagliptin on the reduction in markers were found. Although the exact (36). Therefore a larger-scale prospective carotid IMT may not be overestimated; reason for these inconsistent findings is clinical trial with a longer observation pe- previous reports demonstrated beneficial not clear, several factors could have af- riod is required to assess the CV efficacy of effects of those medications on carotid fected the proinflammatory conditions, DPP-4 inhibitors. On the other hand, we IMT (25,26). Indeed, almost similar find- such as both transient and chronic non- focusedontheeffectsofsitagliptinonthe ings were observed in ANCOVA models atherosclerotic diseases and additional progression of carotid atherosclerosis, not after adjusting for the use of glinides treatment within a relatively short time. CVD, as a first step because of practi- and a-glucosidase inhibitors (data not In addition, a subanalysis of the Diabe- cal constraints, including trial costs and shown). tes Control and Complications Trial/ concern about feasibility in relation to Previous reports demonstrated that Epidemiology of Diabetes Interventions long-term intervention; progression or short-term treatment with DPP-4 inhibitors and Complications (DCCT/EDIC) demon- regression of carotid atherosclerosis decreased the levels of various markers strated that the relationships between se- remains a controversial surrogate for of inflammation and endothelial injury in rum biomarkers of chronic inflammation CV effects (37,38). In addition, the 8 Effects of Sitagliptin on Atherosclerosis Diabetes Care

subjects of the above-mentioned three Although our data suggest a promising Pharmaceutical Co.; Merck Sharp and Dohme trials (29–31) were patients with T2DM effect of sitagliptin on carotid IMT pro- K.K.; Novo Nordisk Pharma Ltd.; Novartis with a history of CVD or at high risk for gression, a large-scale prospective trial Pharmaceuticals Corp.; Co.; Takeda Pharmaceutical Co.; Kissei Pharma- CVD. In comparison, our study enrolled is required to establish the usefulness ceutical Co.; Sumitomo Dainippon Pharma insulin-treated subjects with T2DM who of DPP-4 inhibitors in the primary pre- Co.; Mitsubishi Tanabe Pharma Co.; Kyowa were free from apparent CVD. Consider- vention of CVD in patients with T2DM. Hakko Kirin Co. Ltd.; Eli Lilly & Co.; Pfizer; ing the results of this study, we believe AstraZeneca; and , Inc. and that early and effective intervention with has received research funds from Takeda Phar- maceutical Co.; Merck Sharp and Dohme K.K.; DPP-4 inhibitors before the develop- Acknowledgments. The authors thank all the Mochida Pharmaceutical Co.; Sanofi,Novartis ment of advanced atherosclerosis in pa- clinical staff and members of the committee Pharmaceuticals Corp.; Novo Nordisk Pharma tients without a history of apparent CVD (Yoshimitsu Yamasaki, Nishi-Umeda Clinic for Asian Ltd.; Eli Lilly & Co.; Daiichi Sankyo Co.; fi Medical Collaboration; Kazunori Shimada and & Co. Ltd.; Teijin Pharma Ltd.; Sumitomo is likely to be bene cial in the prevention Hirotoshi Ohmura, Department of Cardiovascular of carotid IMT progression. Dainippon Pharma Co.; Otsuka Pharmaceutical Medicine, Juntendo University Graduate School of Co.; Kissei Pharmaceutical Co.; Mitsubishi Medicine; and Ryota Tanaka, Department of Tanabe Pharma Co.; Ono Pharmaceutical Co.; Neurology, Juntendo University Graduate School Study Limitations AstraZeneca; Astellas Pharma, Inc.; and Kyowa of Medicine) for their assistance with the execution In addition to the limitations described Hakko Kirin Co. Ltd. K.K. has received lecture and completion of the clinical trial. fi above, we used a PROBE design, which fees from Boehringer Ingelheim, Sano , Novo Funding. Financial support for this study was Nordisk Pharma Ltd., Novartis Pharmaceuti- may bias the assessment of outcomes. provided by the Japan Society for Patient Re- cals Corp., Eli Lilly & Co., Takeda Pharma- The reason for using an open-label study ported Outcome research fund from Mitsubishi ceutical Co., Merck Sharp and Dohme K.K., was practical constraints, including trial Tanabe, Ono, and Novo Nordisk. Kowa Pharmaceutical Co. Ltd., Mitsubishi Duality of Interest. T.M. received research Tanabe Pharma Co. and research funds from costs in an investigator-initiated trial. In funds from Merck Sharp and Dohme K.K. and an effort to overcome possible bias, a Sysmex Co. H.Yok. has received lecture fees Takeda Pharmaceutical Co. and has received lec- from Boehringer Ingelheim; Sanofi; Ono Phar- single experienced reader, who was ture fees from AstraZeneca, Boehringer Ingelheim, maceutical Co.; Novo Nordisk Pharma Ltd.; blinded to the clinical characteristics of Eli Lilly & Co., Kowa Pharmaceutical Co. Ltd., Novartis Pharmaceuticals Corp.; Sanwa Kagaku the subjects and type of treatment, Mitsubishi Tanabe Pharma Co., Merck Sharp and Kenkyusho Co.; Daiichi Sankyo Co.; Takeda Dohme K.K., Ono Pharmaceutical Co., and Takeda Pharmaceutical Co.; Merck Sharp and Dohme measured carotid IMT using automated, Pharmaceutical Co. N.Ka. is a staff member of the digital edge-detection technology. Fur- K.K.; Sumitomo Dainippon Pharma Co.; endowed chair (Department of Atherosclerosis Astellas Pharma, Inc.;AstraZeneca;Kyowa ther, all statistical methods and data and Metabolism) established by funds from Hakko Kirin Co. Ltd.; and Kowa Pharmaceutical handling were prespecified under the Kowa Pharmaceutical Co. Ltd. and has re- Co. Ltd. and has received research funds from blinded condition before the database ceived research funds from Merck Sharp and Sanwa Kagaku Kenkyusho Co. M.G. hasreceived Dohme K.K. and lecture fees from Astellas lecture fees from Novartis Pharmaceuticals Corp. was locked. Second, there may have Pharma, Inc.; AstraZeneca; Boehringer Ingel- been measurement errors in IMT as a and Taiho Pharmaceutical Co. Ltd. and has re- heim; Daiichi Sankyo Co.; Sumitomo Dainippon ceived travel fees from Takeda Pharmaceutical result of intersonographer differences Pharma Co.; Co.; Eli Lilly & Co.; Kowa Phar- Co. I.S. has received lecture fees from Astellas among institutions, which were not maceutical Co. Ltd.; Mitsubishi Tanabe Pharma Pharma, Inc.; AstraZeneca; Merck Sharp and evaluated in this study. However, high Co.; Novartis Pharmaceuticals Corp.; Novo Nor- Dohme K.K.; Ono Pharmaceutical Co.; Kyowa disk Pharma Ltd.; Ono Pharmaceutical Co.; reproducibility of the measurement of Hakko Kirin Co. Ltd.; Kowa Pharmaceutical Co. Otsuka Pharmaceutical Co.; Shionogi & Co. fi IMT by each sonographer may reduce Ltd.; Sano ; Sanwa Kagaku Kenkyusho Co.; Daiichi Ltd.; Takeda Pharmaceutical Co.; and Sanofi. Sankyo Co.; Takeda Pharmaceutical Co.; Mitsubishi these errors. Third, the chance of false- T.S. has received lecture fees from Boehringer Tanabe Pharma Co.; Teijin Pharma Ltd.; Eli Lilly positive findings may increase because Ingelheim; Sanofi; Novo Nordisk Pharma Ltd.; Japan; Nippon Boehringer Ingelheim Co.; Novartis the multiplicity of testing for the primary Novartis Pharmaceuticals Corp.; Eli Lilly & Co.; Pharmaceuticals Corp.; Novo Nordisk Pharma end points was not specifically adjusted in Abbott Japan Co.; Takeda Pharmaceutical Co.; Ltd.; Bayer Yakuhin Ltd.; Pfizer Japan, Inc.; Bristol- Sanwa Kagaku Kenkyusho Co.; Mitsubishi our analysis. Finally, it is likely that other Myers Squibb; Mochida Pharmaceutical Co.; Tanabe Pharma Co.; Daiichi Sankyo Co.; Astellas Shionogi & Co. Ltd.; Taisho Toyama Pharmaceu- yet unknown factors and known factors Pharma, Inc.; Ono Pharmaceutical Co.; Merck tical Co.; and Shionogi & Co. Ltd. and has re- such as change in lifestyle, duration and/or Sharp and Dohme K.K.; Shionogi & Co. Ltd.; and ceived research funds from Astellas Pharma, severity of CV risk factors, and family his- Taisho Toyama Pharmaceutical Co. T.On. has re- Inc.; AstraZeneca; Eisai Co.; Merck Sharp and tory of CVD that we did not evaluate in ceived lecture fees from Ono Pharmaceutical Co. Dohme K.K.; Otsuka Pharmaceutical Co.; Ono N.Ku.has received lecture fees from Sanofi and this study can explain the differences in Pharmaceutical Co.; Kaken Pharmaceutical Co.; Novartis Pharmaceuticals Corp. T.Os. has re- Kissei Pharmaceutical Co.; Co. changes in carotid IMT between the two ceived lecture fees from Novo Nordisk, Inc.; Ltd.; Sanofi; Shionogi & Co. Ltd.; Daiichi Sankyo treatment groups. Patients lost to follow- Astellas Pharma, Inc.; Mitsubishi Tanabe Co.; Sumitomo Dainippon Pharma Co.; Takeda up are unlikely to have a major influence Pharma Co.; Sanwa Kagaku Kenkyusho Co.; Pharmaceutical Co.; Mitsubishi Tanabe Pharma on the results because there were no sig- Takeda Pharmaceutical Co.; and Kowa Pharma- Co.; Teijin Pharma Ltd.; Nippon Boehringer ceutical Co. Ltd. and has received research nificant differences in baseline character- Ingelheim Co.; Novartis Pharmaceuticals funds from Novo Nordisk, Inc.; Astellas Corp.; Novo Nordisk Pharma Ltd.; Pfizer istics and AEs between the two groups. Pharma, Inc.; Mitsubishi Tanabe Pharma Co.; Japan, Inc.; Bristol-Myers Squibb; Mochida Sanwa Kagaku Kenkyusho Co.; Kowa Pharma- Pharmaceutical Co.; Eli Lilly Japan; Kowa Pharma- Summary ceutical Co. Ltd.; Novo Nordisk Pharma Ltd.; ceutical Co. Ltd.; and Taisho Toyama Pharmaceu- Compared with the conventional treat- Sumitomo Dainippon Pharma Co.; Eli Lilly & tical Co. H.W. has received lecture fees from Novo ment, sitagliptin treatment attenuated Co.; Taisho Pharmaceutical Co.; GlaxoSmithKline; Nordisk, Inc.; Eli Lilly & Co.; Sanofi; Sumitomo Sumitomo Dainippon Pharma Co.; Astellas Dainippon Pharma Co.; Fujifilm; Bayer Health- the progression of carotid IMT in insu- Pharma US, Inc.; Bayer HealthCare; and Care; Kissei Pharmaceutical Co.; Mochida lin-treated patients with T2DM who AbbVie GK. H.K. has received lecture fees Pharmaceutical Co.; Merck Sharp and Dohme were free of a history of apparent CVD. from Boehringer Ingelheim; Sanofi;Ono K.K.; Takeda Pharmaceutical Co.; Boehringer care.diabetesjournals.org Mita and Associates 9

Ingelheim; Daiichi Sankyo Co.; Ono Pharmaceu- 7. Nemoto M, Tajima N, Kawamori R. Efficacy of with Doppler ultrasound. Med Ultrason 2015; tical Co.; Novartis Pharmaceuticals Corp.; Mitsu- combined use of miglitol in type 2 diabetes 17:167–174 bishi Tanabe Pharma Co.; AstraZeneca; Kyowa patients receiving insulin therapy-placebo- 20. Hodis HN, Mack WJ, Zheng L, et al. Effect of Hakko Kirin Co. Ltd.; Sanwa Kagaku Kenkyusyo controlled double-blind comparative study. peroxisome proliferator-activated receptor Co.; Kowa Pharmaceutical Co. Ltd.; and Astellas Acta Diabetol 2011;48:15–20 gamma agonist treatment on subclinical athero- Pharma, Inc.; has received advisory fees from 8. Katsuno T, Ikeda H, Ida K, Miyagawa J, sclerosis in patients with insulin-requiring Novo Nordisk Pharma Ltd., Inc.; Mochida Phar- Namba M. Add-on therapy with the DPP-4 in- type 2 diabetes. Diabetes Care 2006;29:1545– macuetical Co.; AstraZeneca; Kowa Pharmaceuti- hibitor sitagliptin improves glycemic control 1553 cal Co. Ltd.; Astellas Pharma, Inc.; Sanofi; in insulin-treated Japanese patients with 21. Yasunari E, Takeno K, Funayama H, et al. Boehringer Ingelheim; Merck Sharp and type 2 diabetes mellitus. Endocr J 2013;60: Efficacy of pioglitazone on glycemic con- Dohme K.K.; Mitsubishi Tanabe Pharma Co.; 733–742 trol and carotid intima-media thickness Novartis Pharmaceuticals Corp.; Sumitomo 9. Shimoda S, Iwashita S, Ichimori S, et al. in type 2 diabetes patients with inadequate Dainippon Pharma Co.; Takeda Pharmaceuti- Efficacy and safety of sitagliptin as add-on insulin therapy. J Diabetes Investig 2011;2: – cal Co.; Ono Pharmaceutical Co.; Pfizer; and therapy on glycemic control and blood glu- 56 62 Kowa Pharmaceutical Co. Ltd.; and has received cose fluctuationinJapanese type 2 diabetes 22. Langenfeld MR, Forst T, Hohberg C, et al. Pioglitazone decreases carotid intima-media research funds from Boehringer Ingelheim; Pfizer; subjects ongoing with multiple daily insulin thickness independently of glycemic control in Mochida Pharmaceutical Co.; Sanofi;NovoNor- injections therapy. Endocr J 2013;60:1207– patients with type 2 diabetes mellitus: results disk Pharma Ltd.; Novartis Pharmaceuticals Corp.; 1214 from a controlled randomized study. Circulation Sanwa Kagaku Kenkyusho Co.; Medical 10. Sato S, Saisho Y, Kou K, et al. Efficacy and 2005;111:2525–2531 Corp.; Eli Lilly & Co.; Mitsubishi Tanabe Pharma safety of sitagliptin added to insulin in Japa- 23. Koshiyama H, Shimono D, Kuwamura N, Co.; Daiichi Sankyo Co.; Takeda Pharmaceutical nese patients with type 2 diabetes: the EDIT randomized trial. PLoS One 2015;10: Minamikawa J, Nakamura Y. Rapid communi- Co.; Merck Sharp and Dohme K.K.; Shionogi & cation: inhibitory effect of pioglitazone on Co. Ltd.; Sumitomo Dainippon Pharma Co.; Kissei e0121988 11. Fonseca V, Schweizer A, Albrecht D, Baron carotidarterialwallthicknessintype2diabe- Pharmaceutical Co.; and AstraZeneca. No other tes. J Clin Endocrinol Metab 2001;86:3452– fl MA, Chang I, Dejager S. Addition of vildagliptin potential con icts of interest relevant to this ar- 3456 ticle were reported. to insulin improves glycaemic control in type 2 diabetes. Diabetologia 2007;50:1148–1155 24. Vilsbøll T, Rosenstock J, Yki-Jarvinen¨ H, Author Contributions. T.M. and N.Ka. drafted fi 12. Rosenstock J, Rendell MS, Gross JL, Fleck et al. Ef cacy and safety of sitagliptin when the manuscript. M.G. analyzed research data. added to insulin therapy in patients with T.M.,N.Ka.,T.S.,H.Yos.,T.On.,N.Ku.,T.Os.,H.K., PR, Wilson CA, Mekki Q. Alogliptin added to in- sulin therapy in patients with type 2 diabetes type 2 diabetes. Diabetes Obes Metab 2010; K.K., Y.U., T.Y., K.M., H.Yok., M.T., M.G., I.S. and 12:167–177 reduces HbA(1C) without causing weight gain or H.W. were involved in analysis and interpretation 25. Mita T, Watada H, Shimizu T, et al. Nategli- increased hypoglycaemia. Diabetes Obes Metab of data, reviewed/edited the manuscript, and nide reduces carotid intima-media thickening in fi 2009;11:1145–1152 approved the nal manuscript. The authors meet type 2 diabetic patients under good glycemic 13. Matsubara J, Sugiyama S, Sugamura K, the criteria for authorship recommended by the control. Arterioscler Thromb Vasc Biol 2007; et al. A dipeptidyl peptidase-4 inhibitor, International Committee of Medical Journal Edi- 27:2456–2462 des-fluoro-sitagliptin, improves endothelial func- tors and take full responsibility for all content in 26. Oyama T, Saiki A, Endoh K, et al. Effect of tion and reduces atherosclerotic lesion formation the manuscript and all editorial decisions. I.S. and acarbose, an alpha-glucosidase inhibitor, on se- in apolipoprotein E-deficient mice. J Am Coll H.W. are the guarantors of this work and, as such, rum lipoprotein lipase mass levels and common Cardiol 2012;59:265–276 have full access to all the data in the study and carotid artery intima-media thickness in type 2 14. Ervinna N, Mita T, Yasunari E, et al. take responsibility for the integrity of the data diabetes mellitus treated by sulfonylurea. Anagliptin, a DPP-4 inhibitor, suppresses prolif- and the accuracy of the data analysis. J Atheroscler Thromb 2008;15:154–159 eration of vascular smooth muscles and mono- 27. Tremblay AJ, Lamarche B, Deacon CF, fl References cyte in ammatory reaction and attenuates Weisnagel SJ, Couture P. Effects of sitagliptin fi 1. Haffner SM, Lehto S, Ronnemaa¨ T, Pyor¨ al¨ a¨ atherosclerosis in male apo E-de cient mice. therapy on markers of low-grade inflammation – K, Laakso M. Mortality from coronary heart dis- Endocrinology 2013;154:1260 1270 and cell adhesion molecules in patients with ease in subjects with type 2 diabetes and in non- 15. Shah Z, Kampfrath T, Deiuliis JA, et al. Long- type 2 diabetes. Metabolism 2014;63:1141– diabetic subjects with and without prior term dipeptidyl-peptidase 4 inhibition reduces 1148 fl myocardial infarction. N Engl J Med 1998;339: atherosclerosis and in ammation via effects on 28. Satoh-Asahara N, Sasaki Y, Wada H, et al. A 229–234 monocyte recruitment and chemotaxis. Circula- dipeptidyl peptidase-4 inhibitor, sitagliptin, ex- – 2. Gerstein HC, Miller ME, Byington RP, et al.; tion 2011;124:2338 2349 erts anti-inflammatory effects in type 2 diabetic Action to Control Cardiovascular Risk in Diabe- 16. Mita T, Katakami N, Shiraiwa T, et al. Rationale, patients. Metabolism 2013;62:347–351 tes Study Group. Effects of intensive glucose design, and baseline characteristics of a clin- 29. Hunt KJ, Baker NL, Cleary PA, Klein R, Virella lowering in type 2 diabetes. N Engl J Med ical trial for prevention of atherosclerosis in G, Lopes-Virella MF; DCCT/EDIC Group of Inves- 2008;358:2545–2559 patients with insulin-treated type 2 diabetes tigators. Longitudinal association between en- 3.PatelA,MacMahonS,ChalmersJ,etal.; mellitus using DPP-4 inhibitor: the Sitagliptin dothelial dysfunction, inflammation, and ADVANCE Collaborative Group. Intensive blood Preventive Study of Intima-Media Thickness clotting biomarkers with subclinical atheroscle- glucose control and vascular outcomes in pa- Evaluation (SPIKE). Diabetol Metab Syndr rosis in type 1 diabetes: an evaluation of the tients with type 2 diabetes. N Engl J Med 2014;6:35 DCCT/EDIC cohort. Diabetes Care 2015;38: 2008;358:2560–2572 17. Araki E, Inagaki N, Inoguchi T, et al.; The 1281–1289 4. Duckworth W, Abraira C, Moritz T, et al.; Japanese Diabetes Society. Treatment Guide 30. Balestrieri ML, Rizzo MR, Barbieri M, et al. VADT Investigators. Glucose control and vascu- for Diabetes. Tokyo, Japan, Bunkodo Co., Ltd, Sirtuin 6 expression and inflammatory activity in lar complications in veterans with type 2 diabe- 2014 diabetic atherosclerotic plaques: effects of in- tes. N Engl J Med 2009;360:129–139 18. Yokoyama H, Katakami N, Yamasaki Y. Re- cretin treatment. Diabetes 2015;64:1395–1406 5. Frier BM, Schernthaner G, Heller SR. Hypo- cent advances of intervention to inhibit pro- 31. Arakawa M, Mita T, Azuma K, et al. Inhibi- glycemia and cardiovascular risks. Diabetes Care gression of carotid intima-media thickness in tion of monocyte adhesion to endothelial cells 2011;34(Suppl. 2):S132–S137 patients with type 2 diabetes mellitus. Stroke and attenuation of atherosclerotic lesion 6. Strowig SM, Aviles-Santa´ ML, Raskin P. Com- 2006;37:2420–2427 by a glucagon-like peptide-1 receptor agonist, parison of insulin monotherapy and combina- 19. Plasencia Mart´ınez JM, Garcia Santos JM, exendin-4. Diabetes 2010;59:1030–1037 tion therapy with insulin and metformin or Paredes Martinez ML, Pastor AM. Carotid 32. Nagashima M, Watanabe T, Terasaki M, insulin and troglitazone in type 2 diabetes. Di- intima-media thickness and hemodynamic pa- et al. Native incretins prevent the development abetes Care 2002;25:1691–1698 rameters: reproducibility of manual measurements of atherosclerotic lesions in apolipoprotein E 10 Effects of Sitagliptin on Atherosclerosis Diabetes Care

knockout mice. Diabetologia 2011;54:2649– outcomes in patients with type 2 diabetes 37. Okayama KI, Mita T, Gosho M, et al. Carotid 2659 mellitus. N Engl J Med 2013;369:1317–1326 intima-media thickness progression predicts 33. White WB, Cannon CP, Heller SR, et al.; 35. Green JB, Bethel MA, Armstrong PW, et al.; cardiovascular events in Japanese patients EXAMINE Investigators. Alogliptin after TECOS Study Group. Effect of sitagliptin on car- with type 2 diabetes. Diabetes Res Clin Pract acute coronary syndrome in patients with diovascular outcomes in type 2 diabetes. N Engl 2013;101:286–292 type 2 diabetes. N Engl J Med 2013;369: J Med 2015;373:232–242 38. Lorenz MW, Price JF, Robertson C, et al. 1327–1335 36. Holman RR, Sourij H, Califf RM. Cardiovas- Carotid intima-media thickness progression 34. Scirica BM, Bhatt DL, Braunwald E, et al.; cular outcome trials of glucose-lowering drugs and risk of vascular events in people with dia- SAVOR-TIMI 53 Steering Committee and or strategies in type 2 diabetes. Lancet 2014; betes: results from the PROG-IMT collabora- Investigators. Saxagliptin and cardiovascular 383:2008–2017 tion. Diabetes Care 2015;38:1921–1929