Clinical Care/Education/Nutrition ORIGINAL ARTICLE

Comparison of Vildagliptin and Monotherapy in Patients With A 24-week, double-blind, randomized trial

1 3 JULIO ROSENSTOCK, MD DAVID MILLS, MSC promising new approach to treating 2 3 MICHELLE A. BARON, MD ANJA SCHWEIZER, PHD type 2 diabetes is the augmentation 2 SYLVIE DEJAGER, MD, PHD A of glucagon-like peptide (GLP)-1 receptor signaling by increasing endoge- nous GLP-1 through inhibition of the OBJECTIVE — To compare the efficacy and tolerability of vildagliptin and rosiglitazone dipeptidyl peptidase IV (DPP-4) enzyme during a 24-week treatment in drug-naı¨ve patients with type 2 diabetes. (1). Vildagliptin is a potent and selective DPP-4 inhibitor that improves islet func- RESEARCH DESIGN AND METHODS — This was a double-blind, randomized, ac- tion by increasing both ␣- and ␤-cell tive-controlled, parallel-group, multicenter study of 24-week treatment with vildagliptin (100 ϭ responsiveness to glucose (2,3). Vilda- mg daily, given as equally divided doses; n 519) or rosiglitazone (8 mg daily, given as a gliptin has been shown in 12-week stud- once-daily dose; n ϭ 267). ies to decrease A1C when given as RESULTS — Monotherapy with vildagliptin and rosiglitazone decreased A1C (baseline ϭ monotherapy (4,5) or in combination 8.7%) to a similar extent during the 24-week treatment, with most of the A1C reduction achieved with (6). by weeks 12 and 16, respectively. At end point, vildagliptin was as effective as rosiglitazone, Head-to-head comparison studies re- improving A1C by Ϫ1.1 Ϯ 0.1% (P Ͻ 0.001) and Ϫ1.3 Ϯ 0.1% (P Ͻ 0.001), respectively, cently have been recommended to better meeting the statistical criterion for noninferiority (upper-limit 95% CI for between-treatment establish the efficacy and safety of inves- difference Յ0.4%). Fasting plasma glucose decreased more with rosiglitazone (Ϫ2.3 mmol/l) tigational therapies, such as vildagliptin than with vildagliptin (Ϫ1.3 mmol/l). Body weight did not change in vildagliptin-treated pa- monotherapy, relative to other current Ϫ Ϯ ϩ Ϯ Ͻ tients ( 0.3 0.2 kg) but increased in rosiglitazone-treated patients ( 1.6 0.3 kg, P 0.001 therapies (7). Several classes of drugs are vs. vildagliptin). Relative to rosiglitazone, vildagliptin significantly decreased triglycerides, total approved for the pharmacological treat- cholesterol, and LDL, non-HDL, and total-to-HDL cholesterol (Ϫ9toϪ16%, all P Յ 0.01) but produced a smaller increase in HDL cholesterol (ϩ4 vs. ϩ9%, P ϭ 0.003). The proportion of ment of type 2 diabetes, including the patients experiencing an adverse event was 61.4 vs. 64.0% in patients receiving vildagliptin and (TZDs), rosiglitazone, rosiglitazone, respectively. Only one mild hypoglycemic episode was experienced by one patient and , which are among the in each treatment group, while the incidence of edema was greater with rosiglitazone (4.1%) than most recent additions to the therapeutic vildagliptin (2.1%). armamentarium. Accordingly, the present multicenter, 24-week, double- CONCLUSIONS — Vildagliptin is an effective and well-tolerated treatment option in pa- blind, randomized, controlled clinical tients with type 2 diabetes, demonstrating similar glycemic reductions as rosiglitazone but trial was conducted to compare the effi- without weight gain. cacy and tolerability of monotherapy with vildagliptin (100 mg daily) versus rosiglit- Diabetes Care 30:217–223, 2007 azone (8 mg daily) in drug-naı¨ve patients with type 2 diabetes.

RESEARCH DESIGN AND ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● METHODS — This was a 24-week, From the 1Dallas Diabetes and Endocrine Center, Dallas, Texas; 2Novartis Pharmaceuticals, East Hanover, double-blind, randomized, active- New Jersey; and 3Novartis Pharma AG, Basel, Switzerland. controlled, parallel-group study con- Address correspondence and reprint requests to Julio Rosenstock, MD, Dallas Diabetes and Endocrine ducted at 202 centers in 11 countries in Center at Medical City, 7777 Forest Ln., C-618, Dallas, TX 75230. E-mail: juliorosenstock@ the Americas and Europe. Eligible pa- dallasdiabetes.com. Received for publication 29 August 2006 and accepted in revised form 8 November 2006. tients were randomized to receive vilda- J.R. has served on advisory boards and as a consultant for Pfizer, Roche, Sanofi-Aventis, Novo Nordisk, gliptin 100 mg daily (given as equally MannKind, GlaxoSmithKline, Takeda, Centocor, Johnson & Johnson, Novartis, and ; and has re- divided doses) or rosiglitazone 8 mg daily ceived grant/research support from Merck, Pfizer, Sanofi-Aventis, Novo Nordisk, Bristol-Myers Squibb, Eli (given as a once-daily dose) in a ratio of Lilly, GlaxoSmithKline, Takeda, Novartis, AstraZeneca, Amylin, and Johnson & Johnson. Abbreviations: DPP-4, dipeptidyl peptidase IV; FPG, fasting plasma glucose; GLP, glucagon-like peptide; 2:1. Efficacy and tolerability were as- ITT, intention to treat; TZD, . sessed at weeks 4, 12, 16, and 24 of active A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion treatment. factors for many substances. The study enrolled type 2 diabetic pa- DOI: 10.2337/dc06-1815. Clinical trial reg. no. NCT0009918, clinicaltrials.gov. tients with A1C in the range of 7.5– © 2007 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby 11.0%. These patients had received no marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. pharmacologic treatment for at least 12

DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 217 Vildagliptin versus rosiglitazone monotherapy weeks before screening and no antidia- by either BARC-EU (Ghent, Belgium) for ANCOVA model did not exceed 0.4%. betic agent for Ͼ3 consecutive months at European patients or by Diabetes Diag- For the secondary efficacy variables, tests any time in the past and were considered nostics Laboratory (Columbia, MO) or of superiority were conducted at the two- to be representative of a drug-naı¨ve pop- Covance-US (Indianapolis, IN) for pa- sided significance level of 0.05. In addi- ulation. Male and female patients (nonfer- tients from the Americas. All samples tion, prespecified subanalyses of A1C tile or of childbearing potential using a from any single patient were measured by changes were conducted based on initial medically approved birth control the same laboratory. (baseline) A1C and on BMI category. method), aged 18–80 years, with BMI 22–45 kg/m2 and with fasting plasma Ethics and good clinical practice glucose (FPG) Ͻ15 mmol/l were eligible Analysis populations and data All participants provided written in- to participate. analysis formed consent. The protocol was ap- Patients were excluded if they had a The primary intention-to-treat (ITT) pop- proved by the independent ethics history of type 1 diabetes or secondary ulation consists of all randomized pa- committee/institutional review board at forms of diabetes; acute metabolic diabe- tients who 1) had a screening A1C value each study site, and the study was con- tes complications; myocardial infarction, Ն7.4%, 2) received at least one dose of ducted in accordance with the Declara- unstable angina or coronary artery bypass study medication, and 3) had a baseline as tion of Helsinki, using Good Clinical surgery within the previous 6 months; well as at least one postbaseline A1C mea- Practice. congestive heart failure; liver disease, surement. A total of 89 randomized pa- such as cirrhosis or chronic active hepati- tients were excluded from the primary RESULTS — A total of 786 patients tis; and any contraindications and warn- ITT population for the following reasons: were randomized, and 697 patients com- ings according to the country-specific 4 received no intervention and 13 had no prised the primary ITT population (459 label for rosiglitazone. The following lab- postbaseline A1C measurement; in addi- patients randomized to receive vildaglip- oratory abnormalities were also excluded: tion, 61 patients were inappropriately tin 100 mg daily and 238 patients ran- alanine aminotransferase or aspartate randomized with screening A1C Ͻ7.4%, domized to rosiglitazone 8 mg daily); 782 aminotransferase Ͼ2.5 times the upper and 11 patients had no baseline A1C as- patients comprised the safety population. limit of normal, direct bilirubin Ͼ1.3 sessment due to a systematic error in the Figure 1 summarizes the disposition of times the upper limit of normal, serum measurement of A1C by the U.S. labora- patients from screening through study creatinine levels Ͼ220 ␮mol/l, clinically tory originally used for the study. The end point, and Table 1 reports the demo- significant abnormal thyroid-stimulating U.S. laboratory was subsequently graphic and baseline metabolic character- hormone, or fasting triglycerides Ͼ7.9 changed, and no measurements per- istics of the patients in the primary ITT mmol/l. formed by the initial laboratory are used population. The groups were well bal- A1C, FPG, body weight, and vital in the analyses. The statistical power of anced, with A1C averaging 8.7% and FPG signs were measured at each study visit. the study was preserved by recruitment of averaging 10.3 mmol/l in both treatment Standard hematology and biochemistry additional patients, and all samples from groups. One-third of patients had an A1C laboratory assessments were made at each any single patient were measured by the Ͼ9%. Participants were predominantly visit except on week 16. Fasting lipid pro- same laboratory throughout the study. Caucasian and obese (30% with BMI Ն35 files were measured and electrocardio- The safety population consists of all pa- kg/m2), with a mean age of 54 years and grams were performed at screening and at tients who received at least one dose of the mean disease duration of 2.4 years. More weeks 0, 12, and 24. study drug and had at least one postbase- than 85% of all patients randomized to All adverse events were recorded. line safety assessment. either treatment completed the 24-week Edema was assessed by the investigator as The primary efficacy variable was the study. part of the normal adverse event– change from baseline in A1C at study end reporting process, either as a new occur- point using the last observation carried Efficacy rence or worsening of an existing forward for patients who discontinued Figure 2A depicts the time-course of condition. Patients were provided with early. Secondary efficacy parameters in- mean A1C during the 24-week treatment glucose-monitoring devices and supplies cluded changes in FPG, fasting plasma with vildagliptin 100 mg daily or rosigli- and instructed on their use. Hypoglyce- lipids, and body weight. The efficacy tazone 8 mg daily. Baseline A1C values mia was defined as symptoms suggestive analyses were performed with data from were identical in the two treatment of low blood glucose confirmed by a self- the primary ITT population, which was groups (8.7 Ϯ 0.1%). A1C decreased with monitored blood glucose measurement prespecified as the main efficacy popula- vildagliptin treatment over the entire 24- Ͻ3.1 mmol/l plasma glucose equivalent. tion. Change from baseline in primary week study period, with most of the re- Severe hypoglycemia was defined as any and secondary end points were analyzed duction attained by week 12. episode requiring the assistance of an- using an ANCOVA model, with treatment Rosiglitazone treatment appeared to have other party. and pooled center as the classification a somewhat slower onset of effect, with All laboratory assessments were made variables and baseline as the covariate. A nearly maximum reduction reached at by central laboratories. All assessments, test for the noninferiority of vildagliptin week 16. In the primary ITT population, except A1C, were performed by BARC to rosiglitazone in A1C was carried out the adjusted mean change in A1C from (Bioanalytical Research Corporation). As- through a CI approach. Noninferiority for baseline to study end point was Ϫ1.1 Ϯ says were performed according to stan- A1C was established if the upper limit of 0.1% (P Ͻ 0.001) in patients receiving dardized and validated procedures in the 95% CI for the between-treatment dif- vildagliptin (n ϭ 459) and Ϫ1.3 Ϯ 0.1% accordance with good laboratory prac- ference in the adjusted mean change from (P Ͻ 0.001) in patients receiving rosigli- tice. A1C measurements were performed baseline to end point obtained from the tazone (n ϭ 238). Noninferiority of vilda-

218 DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 Rosenstock and Associates

Figure 1— Disposition of patients from screening through completion.

gliptin 100 mg daily to rosiglitazone 8 mg Table 1—Baseline characteristics of the primary ITT population* daily was established, as the upper limit of the 95% CI for the between-group differ- Vildagliptin 100 mg daily Rosiglitazone 8 mg daily ence in adjusted mean change in A1C (Ϫ0.01 to 0.39) did not exceed the pre- n 459 238 specified noninferiority margin. Age (years) 54.5 Ϯ 11.7 54.2 Ϯ 11.6 The decrease in A1C with either agent Sex was substantially larger in patients with Male 264 (57.5) 137 (57.6) baseline A1C Ͼ9.0%, with mean A1C re- Female 195 (42.5) 101 (42.4) ductions of Ϫ1.8 Ϯ 0.1% (P Ͻ 0.001) Race from a baseline of 10.0% (n ϭ 166) with Caucasian 365 (79.5) 190 (79.8) vildagliptin and of Ϫ1.9 Ϯ 0.2% (P Ͻ Hispanic or Latino 51 (11.1) 29 (12.2) 0.001) from a baseline of 9.9% (n ϭ 88) Black 27 (5.9) 11 (4.6) with rosiglitazone. In the vildagliptin All other 16 (3.5) 8 (3.4) group, patients with BMI Ͻ30 kg/m2 BMI (kg/m2) 32.2 Ϯ 5.7 32.9 Ϯ 6.0 showed a somewhat greater reduction in BMI group (kg/m2) A1C (⌬A1C ϭϪ1.3 Ϯ 0.1%, n ϭ 184) Ͻ30 184 (40.1) 83 (34.9) compared with obese patients with BMI Ն30 275 (59.9) 155 (65.1) Ն30 kg/m2 (⌬A1C ϭϪ1.0 Ϯ 0.1%, n ϭ Ն35 132 (28.8) 76 (31.9) 275). Rosiglitazone, on the other hand, A1C (%) 8.7 Ϯ 1.1 8.7 Ϯ 1.1 was somewhat more efficacious in pa- A1C group (%) tients with BMI Ն30 kg/m2 (⌬A1C ϭ Յ9.0 293 (63.8) 150 (63.0) Ϫ1.4 Ϯ 0.1%, n ϭ 155) than in leaner Ͼ9.0 166 (36.2) 88 (37.0) patients (⌬A1C ϭϪ1.1 Ϯ 0.2%, n ϭ 83). FPG (mmol/l) 10.3 Ϯ 2.7 10.3 Ϯ 2.9 FPG also decreased significantly dur- Disease duration (years) 2.3 Ϯ 3.4 2.7 Ϯ 4.2 ing the 24-week treatment with either Data are means Ϯ SD or n (%). *Baseline characteristics were similar in the randomized population.

DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 219 Vildagliptin versus rosiglitazone monotherapy

Figure 2—A: Mean Ϯ SE A1C during the 24-week treatment with vildagliptin (100 mg daily; Œ) or rosiglitazone (8 mg daily; E) in patients with type 2 diabetes (primary ITT population: vildagliptin, n ϭ 434 at week Ϫ2, n ϭ 397 at week 24; rosiglitazone, n ϭ 221 at week Ϫ2, n ϭ 209 at week 24). B: Adjusted mean change from baseline to end point in body weight in the primary ITT population and in subgroup of patients with BMI Ն35 kg/m2. f, vildagliptin 100 mg daily; , rosiglitazone 8 mg daily. *P Ͻ 0.05; **P Ͻ 0.01; ***P Ͻ 0.001 vs. rosiglitazone. C: Adjusted mean change from baseline to end point in fasting lipid parameters in the primary ITT population. f, vildagliptin 100 mg daily; , rosiglitazone 8 mg daily. *P Ͻ 0.05; **P Ͻ 0.01; ***P Ͻ 0.001 vs. rosiglitazone.

220 DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 Rosenstock and Associates agent. In the primary ITT population, the quent specific adverse events (Ն4% in Although the two agents had similar mean baseline FPG was 10.3 mmol/l in either group) were nasopharyngitis overall efficacy to reduce A1C, the differ- both treatment groups. The FPG reduc- (6.8%), dizziness (6.0%), headache ent mechanism of action of the two agents tion (adjusted mean change) was Ϫ1.3 Ϯ (5.0%), and upper respiratory tract infec- likely underlies several differences 0.1 mmol/l (P Ͻ 0.001) in patients receiv- tion (4.5%). In rosiglitazone-treated pa- noted regarding secondary efficacy end ing vildagliptin and Ϫ2.3 Ϯ 0.2 mmol/l tients, the most common adverse events points as well as in tolerability profiles. (P Ͻ 0.001) in patients receiving rosigli- were nasopharyngitis (7.5%), headache Vildagliptin inhibits the enzyme DPP-4, tazone (P Ͻ 0.001 vs. vildagliptin). (5.2%), dizziness (4.1%), and peripheral causing an increase in active plasma lev- edema (4.1%). The incidence of periph- els of the incretin hormones GLP-1 and Lipids and body weight eral edema with vildagliptin was 2.1%. gastrointestinal polypeptide (3). Vilda- Figure 2 also depicts changes in body Increased body weight was reported as an gliptin has been shown to improve islet weight (B) and fasting lipid parameters adverse event in 0.8% of vildagliptin- function by increasing the ability of (C) after the 24-week treatment with treated patients and in 2.6% of rosiglita- both ␣- and ␤-cells to sense and re- vildagliptin 100 mg daily or rosiglitazone zone-treated patients. One patient in each spond appropriately to glucose (2,8). 8 mg daily in the primary ITT population. group reported one mild hypoglycemic These effects are thought to be mediated Body weight at baseline averaged 91.2 Ϯ event, and no serious hypoglycemic by GLP-1 (9). In contrast, the TZDs tar- 0.9 kg in the vildagliptin group and events occurred in either group. get resistance acting by activa- 93.1 Ϯ 1.3 kg in the rosiglitazone group. The proportion of patients experienc- tion of peroxisome proliferator– Body weight did not change during 24- ing any serious adverse event in the two activated ␥ receptors, which results in week treatment with vildagliptin but in- treatment groups was comparable (2.9 vs. enhanced peripheral and hepatic insu- creased significantly during rosiglitazone 3.0%), and no specific serious adverse lin action (10). Furthermore, TZDs monotherapy. The between-treatment event was reported by more than one pa- stimulate differentiation of preadipo- difference in body weight was Ϫ1.9 Ϯ 0.3 tient within a treatment group. The fre- cytes into new, small, and highly insu- kg (P Ͻ 0.001). In addition, a post hoc quency of discontinuations due to lin-sensitive fat cells (10). This analysis indicated that in the more se- adverse events was also similar in the promotes storage of free fatty acids in verely obese population (BMI Ն35 kg/ vildagliptin (2.9%) and the rosiglitazone adipose tissue, thus relieving the liver m2; mean body weight of ϳ111 kg; n ϭ (3.4%) groups. There was one death dur- and muscle from lipotoxicity and re- 208), a larger decrease in body weight was ing the study. This was a 70-year-old male ducing gluconeogenesis (11). The de- seen with vildagliptin, while the increase subject randomized to vildagliptin who crease in FPG in vildagliptin-treated seen with rosiglitazone monotherapy was died from postsurgical complications. patients seen in the present study was similar to the overall cohort. The With the exception of a slightly significantly less than that in patients between-treatment difference in body higher proportion of patients with nota- receiving rosiglitazone, but the A1C im- weight in this subpopulation was Ϫ2.8 Ϯ ble hematocrit and hemoglobin abnor- provements were similar, suggesting a 0.6 kg (P Ͻ 0.001). malities in the rosiglitazone group, there more pronounced effect of vildagliptin In the primary ITT population, fast- were no major changes from baseline to on plasma glucose levels in the post- ing lipid levels were similar in the two end point nor were there any between- prandial period and throughout the treatment groups at baseline, averaging treatment differences observed for any day. 2.3 mmol/l for triglycerides, 5.3 mmol/l laboratory parameter or vital signs. The Many effective antidiabetes agents for total cholesterol, 3.1 mmol/l for LDL, frequency of treatment-emergent electro- lead to some weight gain as a result of 1.2 mmol/l for HDL, and 4.1 mmol/l for cardiogram abnormalities was low and improved glycemic control (12), and this non-HDL cholesterol in the combined co- comparable in the two treatment groups. is a particular limitation of (13) and po- hort, with a total-to-HDL cholesterol ratio tential safety concern about TZDs, due to of 4.7. Relative to rosiglitazone, vildaglip- CONCLUSIONS — This study dem- their tendency to cause fluid retention tin produced significant decreases in tri- onstrated that in patients representative and edema (14). In this study, the in- glycerides (Ϫ9%, P ϭ 0.010) and total of a drug-naı¨ve population, vildagliptin crease in body weight and incidence of (Ϫ14%, P Ͻ 0.001), LDL (Ϫ16%, P Ͻ was well tolerated and caused no weight edema in patients receiving rosiglitazone 0.001), and non-HDL cholesterol gain despite a significant and clinically that was observed is consistent with pre- (Ϫ16%, P Ͻ 0.001) but less improve- meaningful decrease from baseline in vious reports, whereas vildagliptin ment in HDL cholesterol (ϩ4 vs. ϩ9% A1C that was similar to that with rosigli- achieved a comparable improvement in from baseline, P ϭ 0.003 for tazone. As expected, both vildagliptin and glycemic control with no weight gain and between-group difference). Relative to rosiglitazone produced more substantial a low incidence of edema. The ability of rosiglitazone, vildagliptin decreased to- reductions in A1C in the subgroup of pa- vildagliptin to improve glycemic control tal-to-HDL cholesterol by 9.1 Ϯ 1.9% tients with a high baseline level, and as in without weight gain needs to be further (P Ͻ 0.0001). the whole cohort, the improvement in clarified mechanistically. glycemic control was similar in patients In the present study, relative to ros- Tolerability with high baseline A1C receiving vilda- iglitazone, vildagliptin treatment was as- During the 24-week treatment, one or gliptin (⌬ϭϪ1.8%) or rosiglitazone sociated with a significant improvement more adverse event was reported by (⌬ϭϪ1.9%). Vildagliptin appeared to in triglycerides; total, LDL, and non-HDL 61.4% of patients receiving vildagliptin be slightly more effective than rosiglita- cholesterol; and, importantly, the total- 100 mg daily and by 64.0% of patients zone in patients with BMI Ͻ30 kg/m2, to-HDL cholesterol ratio. The changes in receiving rosiglitazone 8 mg daily. In pa- and rosiglitazone was slightly more effec- fasting lipids seen in rosiglitazone-treated tients receiving vildagliptin, the most fre- tive in obese patients (BMI Ն30 kg/m2). patients were consistent with those re-

DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 221 Vildagliptin versus rosiglitazone monotherapy ported in previous studies (15–17). The Nevalainen, Dr. Pentti Jarvinen. France: Egelhof, Dr. John Thomson, Dr. John mechanism underlying the improvement Dr. Christian Faugere, Dr. Guillaume Agaiby, Dr. Stephen Hippler, Dr. Thomas in lipid profile seen in vildagliptin-treated Dantin, Dr. Jean Francois Ravaut, Dr. Higgins, Dr. Thomas Wade, Dr. Terence patients is unknown but could reflect a Ge´rard Lalanne, Dr. Philippe Blanchard. Hart, Dr. Andrew Slaski, Dr. John McGet- chronic improvement in postprandial lip- Germany: Dr. Manfred Reiss, Dr. Med. tigan, Dr. Kimy Charani, Dr. Natalie She- ids, as suggested by a recent report that Michael Ziegler, Dr. Reinhold Schneider, monsky, Dr. John Wadleigh, Dr. Michael found decreased postprandial lipemia Dr. Wolfgang Boerner, Dr. Hans Seibert, McAdoo, Dr. Pedro Velasquez-Mieyer, primarily through a reduction in intesti- Dr. Med. Juergen Matthes, Dr. Doris Boe- Dr. Arthur Pitterman, Dr. Alan Wynne, nally derived apolipoprotein B-48– hme, Dr. Matthias Roevenich, Dr. Hans- Dr. S. Archer, Dr. Howard Ellison, Dr. containing particles after a 4-week Joachim Herrmann, Dr. Frank Klein, Dr. Anicia Villafria, Dr. A. Clifton Cage, Dr. treatment with vildagliptin (8). Med. Arthur Sterzing, Dr. Med. Edmond David Damian, Dr. John Sibille, Dr. Peter- With the exception of a higher inci- Homsy, Dr. Daniel Ayasse, Dr. Gerhard man Prosser, Dr. Richard Ferreras, Dr. dence of edema in rosiglitazone-treated Krehan, Dr. Karin Todoroff, Dr. Med. Ru- Ruth Smothers, Dr. W. David Clark, Dr. patients, the two agents were similarly dolf Fuchs. Hungary: Dr. Istvan Witt- Mark Runde, Dr. Chet Monder, Dr. Mat- well tolerated in this 24-week study of mann, Dr. Janos Penzes, Dr. Peter Torzsa, thew Acampora, Dr. Paul Fiacco, Dr. vildagliptin 100 mg daily versus rosiglita- Dr. Jozsef Rinfel, Dr. Tibor Fulop, Dr. Ju- Brian Kauth, Dr. Ronald Sockolov, Dr. zone 8 mg daily, and there was a very low dit Simon, Dr. Sandor Palla, Dr. Horten- Gregory Smith, Dr. John Devlin, Dr. incidence of hypoglycemia. zia Karolyi, Dr. Gyula Neuwirth, Dr. Daniel Sheerer, Dr. Lawrence Levinson, In conclusion, similar A1C efficacy Gyozo Vandorfi. Spain: Dr. Iskra Ligu¨- Dr. Eli Ipp, Dr. David Mansfield, Dr. Paul can be achieved using the DPP-4 inhibitor erre, Dr. Manuel Munoz, Dr. Carlos Al- Davis, Dr. Michael Campolo. vildagliptin or a TZD as monotherapy in mendro, Dr. Belen Fraile, Dr. Jose drug-naı¨ve patients with type 2 diabetes. Gonzalez Clemente, Dr. Magdalena Her- Vildagliptin is well tolerated, and, despite nandez, Dr. Fernando Quirce, Dr. Nidia References the improvement in glycemic control, it Ruiz, Dr. Ildefonso Espinosa, Dr. Luis De 1. Deacon CF, Holst JJ: Dipeptidyl peptidase does not cause weight gain, which is an Teresa. Sweden: Prof. Ulf Adamsson, Dr. IV inhibitors: a promising new therapeu- important consideration in the decision- Arvo Hanni, Dr. Mona Landin-Olsson, tic approach for the management of type 2 making process for selecting first-line Dr. Bo Polhem, Prof. Stephan Rossner, diabetes. Int J Biochem Cell Biol 38:831– 844, 2006 therapy in type 2 diabetes. Dr. Anders Sjoberg, Dr. Mats Dahl, Dr. 2. Ahre´n B, Landin-Olsson M, Jansson P-A, Gunnar Stromblad, Dr. Anders Nilsson, Svenson M, Holmes D, Schweizer A: In- Dr. Anders Norrby. U.S.: Dr. Donald hibition of dipeptidyl peptidase-4 reduces Acknowledgments— This study was funded Huffman, Dr. David Colan, Dr. Ronald glycemia, sustains insulin levels and re- by Novartis Pharmaceuticals. A list of investi- Graf, Dr. Richard Cherlin, Dr. Terry Pol- duces glucagon levels in type 2 diabetes. gators is provided in the APPENDIX. ing, Dr. Kashif Latif, Dr. Robert Hippert, J Clin Endocrinol Metab 89:2078–2084, The authors gratefully acknowledge the in- Dr. Anna Jackson, Dr. Terence Isakov, Dr. 2004 vestigators and staff at the 202 participating Usah Lilavivat, Dr. Neal Shealy, Dr. Jef- 3. Mari A, Sallas WM, He YL, Watson C, sites and the editorial assistance of and helpful Ligueros-Saylan M, Dunning BE, Deacon discussion with Beth Dunning Lower, PhD, frey Newman, Dr. Monica Perlman, Dr. Vin Tangpricha, Dr. Gerard Stanley, Dr. CF, Holst JJ, Foley JE: Vildagliptin, a who is a subcontractor for Novartis. dipeptidyl peptidase-IV inhibitor, im- Paul Dudley, Dr. James Lehman, Dr. proves model-assessed ␤-cell function in Christopher Case, Dr. George Handey, patients with type 2 diabetes. J Clin Endo- APPENDIX Dr. Gobivenkata Balaji, Dr. Francis Yemo- crinol Metab 90:4888–4894, 2005 fio, Dr. Arthur Mollen, Dr. Flor Geola, Dr. 4. Ristic S, Byiers S, Foley J, Holmes D: Im- List of investigators Eric Klein, Dr. J. Forsythe, Dr. Raymond proved glycaemic control with dipeptidyl Argentina: Dr. Alfredo Lozada, Dr. Adri- Grenfell Jr., Dr. Abraham Areephanthu, peptidase-4 inhibition in patients with ana Osorio, Dr. Gustavo Frechtel, Dr. Dr. F. Lester, Dr. Larry Stonesifer, Dr. type 2 diabetes: vildagliptin (LAF237) Marcelo Rusculleda, Dr. Maria Isabel Kly- Kristine Bordenave, Dr. George Ryckman, dose response. Diabetes Obes Metab ver de Saleme, Dr. Guillermo Marcucci, Dr. Thomas Moretto, Dr. Jay Shubrook, 7:692–698, 2005 Dr. Jorge Gomez. Austria: Prim. Doz. Dr. Dr. Danny Sugimoto, Dr. Othman 5. Pratley RE, Jauffret-Kamel S, Galbreath E, Holmes D: Twelve-week monotherapy Attila Dunky, Dr. Franz Winkler, Prof. Shemisa, Dr. Joseph Aloi, Dr. William with the DPP-4 inhibitor vildagliptin im- Hermann Toplak, Univ. Doz. Dr. Rai- Long Jr., Dr. Frank Civitarese, Dr. Rich- proves glycemic control in subjects with mund Weitgasser, Prim. Univ. Prof. Ru- ard Cook, Dr. Matthew Portz, Dr. An- type 2 diabetes. Horm Metab Res 38:423– dolf Prager. Brazil: Dr. Maria Cerqueira, thony Bartkowiak, Dr. Steven Levine, Dr. 428, 2006 Dr. Freddy Eliaschewitz, Dr. Maria Rajeev Kumar Jain, Dr. David Wright, Dr. 6. Ahre´n B, Gomis R, Standl E, Mills D, Zanella, Dr. Marcos Tambascia. Canada: Carl Griffin, Dr. Alan Garber, Dr. Sher- Schweizer A: Twelve- and 52-week efficacy Dr. Chantal Godin, Dr. Kevin Saunders, wyn Schwartz, Dr. Harold Fields, Dr. of the dipeptidyl peptidase IV inhibitor Dr. Dennis O’Keefe, Dr. Errol Raff, Dr. Jerry Mitchell, Dr. Jayaram Naidu, LAF237 in metformin-treated patients with Gottesman, Dr. Jean-Francois Yale, Dr. Dr. Kenneth Hershon, Dr. Gregory type 2 diabetes. Diabetes Care 27:2874– Ronnie Aronson, Dr. Yaw Twum-Barima, Gottschlich, Dr. Curtis Brown, Dr. Berto 2880, 2004 7. Nathan DM, Buse JB, Davidson MB, Heine Dr. Andre Nadeau, Dr. Stuart Ross, Dr. Zamora, Dr. Deborah Thompson, Dr. RJ, Holman RR, Sherwin R, Zinman B: Ronald Goldenberg. Finland: Dr. Jaakko Scott Yates, Dr. Angela Adelizzi, Dr. Mi- Management of hyperglycemia in type 2 Tuomilehto, Dr. Johan Eriksson, Prof. chael Lai, Dr. Garland Thorn Jr., Dr. Tim- diabetes: a consensus algorithm for the Sirkka Keinanen-Kiukaanniemi, Dr. Veli othy Howard, Dr. Larry Cowan, Dr. initiation and adjustment of therapy: a Sillanpaa, Dr. Jorma Lahtela, Dr. Pasi Deborah Cole-Sedivy, Dr. Richard consensus statement from the American

222 DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 Rosenstock and Associates

Diabetes Association and the European liver: decreased gluconeogenesis in pa- 15. Goldberg RB, Kendall DM, Deeg MA, Association for the Study of Diabetes. Di- tients with type 2 diabetes. J Clin Endocri- Buse JB, Zagar AJ, Pinaire JA, Tan MH, abetes Care 29:1963–1972, 2006 nol Metab 91:806–812, 2006 Khan MA, Perez AT, Jacober SJ: A com- 8. Matikainen N, Ma¨ntta¨ri S, Schweizer A, 12. Comi RJ: Treatment of type 2 diabetes parison of lipid and glycemic effects of Ulvestad A, Mills D, Dunning BE, Foley mellitus: a weighty enigma. Ann Intern pioglitazone and rosiglitazone in patients JE, Taskinen M-R: Vildagliptin therapy re- Med 143:609–610, 2005 with type 2 diabetes and dyslipidemia. Di- duces postprandial intestinal triglyceride- 13. Fonseca V: Effect of thiazolidinediones on abetes Care 28:1547–1554, 2005 rich lipoprotein particles in patients with body weight in patients with diabetes 16. Lebovitz HE, Dole JF, Patwardhan R, Rap- type 2 diabetes. Diabetologia 49:2049– mellitus. Am J Med 115 (Suppl. 8A):42S– paport EB, Freed MI: Rosiglitazone 48S, 2003 2057, 2006 monotherapy is effective in patients with 9. Holst JJ, Deacon CF: Glucagon-like pep- 14. Nesto RW, Bell D, Bonow RO, Fonseca V, type 2 diabetes. J Clin Endocrinol Metab tide-1 mediates the therapeutic actions of Grundy SM, Horton ES, Le WM, Porte D, DPP-IV inhibitors. Diabetologia 48:612– Semenkovich CF, Smith S, Young LH, 86:280–288, 2001 615, 2005 Kahn R: Thiazolidinedione use, fluid re- 17. Phillips LS, Grunberger G, Miller E, Pat- 10. Yki-Jarvinen H: Thiazolidinediones. tention, and congestive heart failure: a wardhan R, Rappaport EB, Salzman A: N Engl J Med 351:1106–1118, 2004 consensus statement from the American Once- and twice-daily dosing with rosigli- 11. Gastaldelli A, Miyazaki Y, Pettiti M, San- Heart Association and American Diabetes tazone improves glycemic control in pa- tini E, Ciociaro D, DeFronzo RA, Ferran- Association. Diabetes Care 27:256–263, tients with type 2 diabetes. Diabetes Care nini E: The effect of rosiglitazone on the 2004 24:308–315, 2001

DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 223