Effects of Adding Linagliptin to Basal Insulin Regimen for Inadequately Controlled Type 2 Diabetes a $52-Week Randomized, Double-Blind Study

Clinical Care/Education/Nutrition/Psychosocial Research ORIGINAL ARTICLE

Effects of Adding Linagliptin to Basal Insulin Regimen for Inadequately Controlled Type 2 Diabetes A $52-week randomized, double-blind study

1 4 HANNELE YKI-JÄRVINEN, MD, PHD SUDIPTA BHATTACHARYA, MSJ addition, patients and physicians are of- 2 5 JULIO ROSENSTOCK, MD SANDRA THIEMANN, PHD 3 6 ten reluctant to titrate insulin because of SANTIAGO DURAN -GARCIA, MD, PHD SANJAY PATEL, MB CHB 4 5 the fear of hypoglycemia. An alternative SABINE PINNETTI, MD HANS-JUERGEN WOERLE, MD option to improve glycemic control in basal insulin–treated patients is to add an- d ﬁ other OAD to metformin, such as a dipep- OBJECTIVE To evaluate the ef cacy and long-term safety of linagliptin added to basal tidyl peptidase 4 (DPP-4) inhibitor.DPP-4 insulins in type 2 diabetes inadequately controlled on basal insulin with or without oral agents. inhibitors are an attractive option, because

RESEARCH DESIGN AND METHODSdA total of 1,261 patients (HbA1c $7.0% [53 they lower postprandial glucose concentra- mmol/mol] to #10.0% [86 mmol/mol]) on basal insulin alone or combined with metformin and/ tions more than FPG levels (5) and do not or pioglitazone, were randomized (1:1) to double-blind treatment with linagliptin 5 mg once appear to increase the risk of hypoglycemia daily or placebo for $52 weeks. The basal insulin dose was kept unchanged for 24 weeks but or weight gain. ’ could thereafter be titrated according to fasting plasma glucose levels at the investigators dis- Linagliptin is a DPP-4 inhibitor that is cretion. The primary end point was the mean change in HbA1c from baseline to week 24. The mainly excreted through the bile and gut, safety analysis incorporated data up to a maximum of 110 weeks. unlike other DPP-4 inhibitors, which pri-

RESULTSdAt week 24, HbA1c changed from a baseline of 8.3% (67 mmol/mol) by 20.6% marily undergo renal excretion. Dose ad- (26.6 mmol/mol) and by 0.1% (1.1 mmol/mol) with linagliptin and placebo, respectively (treat- justment is not required with linagliptin, ment difference 20.65% [95% CI 20.74 to 20.55] (27.1 mmol/mol); P , 0.0001). Despite the regardless of the degree of renal or hepatic option to uptitrate basal insulin, it was adjusted only slightly upward (week 52, linagliptin 2.6 IU/ impairment (6,7). This is beneficial in pa- , day, placebo 4.2 IU/day; P 0.003), resulting in no further HbA1c improvements. Frequencies tients who require insulin treatment, be- of hypoglycemia (week 24, linagliptin 22.0%, placebo 23.2%; treatment end, linagliptin cause they frequently have more advanced 31.4%, placebo 32.9%) and adverse events (linagliptin 78.4%, placebo 81.4%) were similar disease, may be older, or may already have between groups. Mean body weight remained unchanged (week 52, linagliptin 20.30 kg, placebo 20.04 kg). renal impairment. The potentially comple- mentary effects of DPP-4 inhibitors and CONCLUSIONSdLinagliptin added to basal insulin therapy significantly improved glyce- basal insulin on postprandial glucose and mic control relative to placebo without increasing hypoglycemia or body weight. FPG, respectively, coupled with the low risk of weight gain or hypoglycemia associated with DPP-4 inhibitors, provide a clinical rationale for adding a DPP-4 inhibitor ecause of the progressive deteriora- treat-to-target studies showed successful specifically to basal insulin therapy (with b Btion of pancreatic -cell function in improvement of HbA1c close to 7% (53 or without additional OADs). DPP-4 inhib- type 2 diabetes, many patients will mmol/mol) when a structured insulin ti- itors were shown to be effective and safe eventually require the addition of a basal tration regimen was systematically imple- when added to insulin therapy in type 2 di- insulin regimen to their existing oral anti- mented in a clinical research setting (2,3). abetes; however, those studies were con- hyperglycemic drugs (OADs) (1). Dosage Achieving the desired FPG level is diffi- ducted in heterogeneous populations on of basal insulin is usually adjusted (ti- cult in clinical practice, however, mainly multiple different insulin regimens without trated) to achieve a fasting plasma glucose because of treatment inertia and failure to the option to uptitrate insulin doses (8–13). (FPG) level of 5.5–6.1 mmol/L. Most empower the patient to self-titrate (4). In The aim of this study was to investi- gate the efficacy and safety of linagliptin ccccccccccccccccccccccccccccccccccccccccccccccccc 5mgoncedailyfor24weeksasadd-on From the 1Department of Medicine, University of Helsinki and Minerva Foundation Institute for Medical therapy in a homogenous population of Research, Helsinki, Finland; the 2Dallas Diabetes and Endocrine Center at Medical City, Dallas, Texas; the basal insulin–treated patients with type 2 3Valme Hospital Medical School, University of Sevilla, Seville, Spain; 4Boehringer Ingelheim Pharma GmbH & diabetes and inadequate glycemic con- 5 Co. KG, Biberach, Germany; Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany; and trol. As part of the unique study design, 6Boehringer Ingelheim Ltd., Bracknell, U.K. Corresponding author: Hannele Yki-Järvinen, [email protected].fi. after the 24-week period, free insulin ti- Received 31 December 2012 and accepted 9 July 2013. tration was allowed up to at least week 52 DOI: 10.2337/dc12-2718. Clinical trial reg. no. NCT 00954447, clinicaltrials.gov. at the investigators’ discretion. This deter- This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10 mined longer term safety of linagliptin and .2337/dc12-2718/-/DC1. © 2013 by the American Diabetes Association. Readers may use this article as long as the work is properly tested whether changes in glycemic control cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ were sustained or necessitated an increase licenses/by-nc-nd/3.0/ for details. in basal insulin dose. care.diabetesjournals.org DIABETES CARE 1 Diabetes Care Publish Ahead of Print, published online September 23, 2013 Linagliptin as add-on to basal insulin

RESEARCH DESIGN AND for Good Clinical Practice. All patients and mean change in body weight to the METHODS gave written, informed consent before end of treatment. participation. Safety end points included the fre- Study design and patients Patients underwent a 2-week, open- quency and intensity of adverse events This randomized, double-blind, placebo- label placebo run-in period to confirm (AEs), including hypoglycemia and clin- controlled, phase III study was conducted their eligibility after the initial screening ically relevant new or worsening findings in 167 centers in 19 countries (Argentina, and to exclude those who were nonad- in physical examination, 12-lead electro- Belgium, Brazil, Canada, Czech Republic, herent. After this placebo run-in period, cardiogram, vital signs, lipid parameters, Finland, Germany, Greece, Italy, Korea, eligible patients were randomly as- and clinical laboratory assessments. An Mexico, the Netherlands, Norway, Peru, signed (1:1 ratio) to receive double-blind independent external adjudication com- Russia, Slovakia, Spain, Taiwan, and linagliptin 5 mg once-daily or placebo in mittee reviewed treatment-emergent fatal the U.S.). addition to continued basal insulin for at events and suspected events of stroke or Patients were eligible if they were least 52 weeks. Treatment assignment cardiac ischemia (including myocardial $18 years of age with a diagnosis of was determined by computer-generated infarction), hospitalization for heart fail- type 2 diabetes, had inadequate glycemic random sequence with an interactive ure, stent thrombosis, and revasculariza- control (HbA1c $7.0% [53 mmol/mol] to voice response system. Randomization tion procedures. Follow-up for all AEs, #10.0% [86 mmol/mol]), had a BMI of was stratified by HbA1c (,8.5% [69 including those persisting after a patient #45 kg/m2, and were receiving treatment mmol/mol] vs. $8.5% [69 mmol/mol]), had completed (or withdrawn prema- with basal insulin, alone or in combina- renal function (estimated glomerular fil- turely from) the trial, continued until tion with metformin and/or pioglitazone, tration rate [eGFR]), and concomitant the event had resolved or been sufficiently for $12 weeks. Acceptable basal insulins use of OADs (metformin only, pioglita- characterized. were insulin glargine, insulin detemir, zone only, metformin and pioglitazone, and neutral protamine Hagedorn insulin. or none). Because the patients who were Statistical analyses The total prescribed insulin dose must randomized early in the trial were treated Allowing for SD of change in HbA1c from not have changed by .10% of the base- until the study’s end, the maximal possible baseline of 1.2% (13.1 mmol/mol), 284 line value during the 12 weeks before treatment duration was 110 weeks. Dur- patients per treatment group were suffi- randomization. ing the first 24 weeks of treatment, the cient to achieve 93% power to detect a Patients were ineligible if they had doses of basal insulin (within 10% of base- 0.35% (3.8 mmol/mol) difference be- uncontrolled fasting hyperglycemia (glu- line dose) and OADs were kept un- tween groups in change in HbA1c from cose .13.3 mmol/L during placebo run- changed. After week 24, adjustments to baseline to week 24. The larger sample in); a myocardial infarction, stroke, or the dose of basal insulin (but not OADs) size of 600 patients in each group allowed transient ischemic attack within 6 months were allowed according to the medical adequate exposure data to be collected for before informed consent; impaired he- judgment of the investigator, with a treat- the regulatory requirement to detect rare patic function (either alanine transami- ment target for FPG of 6.1 mmol/L. cardiovascular events across the entire nase, aspartate transaminase, or alkaline Rescue therapy could be initiated linagliptin program. phosphate .3 times the upper limit of during randomized treatment if a patient The primary end point was evaluated normal); previous gastric bypass surgery; met the following criteria: confirmed FPG with ANCOVA at the level of a = 0.025 or any medical history of cancer (except (after overnight fast) .13.3 mmol/L dur- (one-sided). The statistical model included basalcellcarcinoma)inthe5yearsbefore ing the first 12 weeks, FPG .11.1 mmol/L “treatment,”“concomitant OADs,” and screening. Further exclusion criteria in- from weeks 12–24, or FPG .10.0 mmol/L “baseline renal function impairment cate- . ” fi fi cluded hypersensitivity or allergy to the or HbA1c 8.0% (64 mmol/mol) after gory as xed classi cation effects and investigational products; contraindica- week 24. For initiation of rescue medica- “baseline HbA1c” as linear covariate. This tions to metformin or pioglitazone; treat- tion, these criteria had to be confirmed by analysis was performed on the full anal- ment with rosiglitazone, sulfonylureas, two measurements on separate days. Pa- ysis set (FAS), comprising all random- glucagon-like peptide 1 analogs, DPP-4 tients were withdrawn from the trial if the ized patients treated with at least one inhibitors, or antiobesity drugs within FPG remained above this threshold despite dose of study medication, with a baseline the 3 months before informed consent; a rescue therapy. HbA1c measurement and at least one on- history of alcohol or drug abuse in the treatment HbA1c measurement within previous 3 months; and current treatment End points and assessments the first 24 weeks of double-blind treat- with systemic steroids or change in dos- The primary efficacy end point was the ment. An approach of last observation age of thyroid hormones within 6 weeks change from baseline in HbA1c after 24 carried forward (LOCF) was used to re- before informed consent. Premenopausal weeks of treatment. Secondary end points place missing data. women who were nursing, pregnant, or included changes from baseline in HbA1c Secondary end points were evaluated not practicing an acceptable method of and FPG with time, change from baseline in the FAS with an ANCOVA model with birth control were also ineligible. in FPG after 52 weeks of treatment, the LOCF. Changes in FPG with time were The trial protocol was approved by proportion of patients achieving HbA1c analyzed for the FAS (observed case set the independent ethics committees or ,7% (53 mmol/mol), the proportion of [OC]; i.e., patients with available data) by institutional review boards of all partici- patients achieving $0.5% (5.5 mmol/mol) means of descriptive statistics. The impact pating centers. The study was carried out reduction in HbA1c, and the change from of treatment on the use of rescue medica- according to the principles of the Decla- baseline in mean basal insulin dose after tion was assessed by means of logistic re- ration of Helsinki and the International 52 weeks of treatment. Other end points gression, and the time to first use of rescue Conference on Harmonization Guideline included the use of rescue medication therapy was evaluated by Kaplan-Meier

2 DIABETES CARE care.diabetesjournals.org Yki-Järvinen and Associates analysis. The impact of treatment on the (linagliptin 4.0%, placebo 5.2%) and refusal (95% CI 20.9 to 20.4; P , 0.001). Dur- occurrence of hypoglycemia was investi- to continue study medication (linagliptin ing the period when insulin titration was gated by means of logistic regression and 3.3%, placebo 4.1%). Mean exposures to allowed and recommended, improvement Kaplan-Meier analysis. Safety end points study medication were 435.5 days in the in FPG was sustained in the linagliptin were evaluated for the treated set (all linagliptin group and 422.4 days in the group, whereas FPG decreased in the pla- patients who were treated with at least placebo group (medians 444 and 448 cebo group in parallel with a small in- one dose of study medication) by means days, respectively). crease in basal insulin dose (Fig. 2B). By of descriptive statistics. AEs were de- Demographic and clinical characteris- week 52, the mean changes in FPG from scribed according to the Medical Dictio- tics were similar between treatment groups baseline were similar between the groups nary for Drug Regulatory Affairs (version at baseline and are presented in Table 1. (linagliptin baseline 7.8 mmol/L, 52-week 14.0). Hypoglycemia was analyzed by 7.6 mmol/L, change from baseline 20.2 three levels of intensity: plasma glucose Efficacy outcomes [0.2] mmol/L; placebo baseline 7.8 mmol/L, #4 mmol/L accompanied by typical Changes in HbA1c and FPG. Linagliptin 52-week 7.6 mmol/L, change from baseline symptoms of hypoglycemia, plasma glu- was superior to placebo in reducing 20.3 [0.2] mmol/L). The FPG target of , , cose 3 mmol/L accompanied by typical HbA1c after 24 weeks (Supplementary 6.1 mmol/L was achieved by 24.6% of symptoms of hypoglycemia but without Table 1). The adjusted mean (SE) change patients in the linagliptin group and need for external assistance, and severe in HbA1c frombaselineatweek24for 19.1% of the placebo group. Of these pa- hypoglycemia requiring the assistance of linagliptin was 20.58% (0.08) (26.3 [0.9] tients, 29.6% and 38.1% experienced hy- another person to actively administer car- mmol/mol), compared with 0.07% (0.08) poglycemia in the linagliptin and placebo bohydrate, glucagon, or other resuscita- (0.8 [0.9] mmol/mol) for placebo, result- groups, respectively. tive actions. ing in a placebo-adjusted mean change Among patients with baseline HbA1c in HbA1c from baseline of 20.65% (95% $7.0% (53 mmol/mol), after 52 weeks RESULTS CI 20.74 to 20.55) (27.1 mmol/mol; an HbA1c ,7.0% (53 mmol/mol) was P , 0.0001). After 52 weeks, the adjusted achieved by 16% and 7% of the linagliptin , Patient disposition, demographics, mean changes in HbA1c from baseline were and placebo groups, respectively (P and baseline clinical characteristics 20.48% (0.08) (25.2 [0.9] mmol/mol) for 0.001). Compared with the placebo This study was conducted between Au- linagliptin and 0.05% (0.08) (0.5 [0.9] group, patients in the linagliptin group gust 2009 and September 2011. A total of mmol/mol) for placebo (placebo-adjusted were more likely to have a decrease in 1,261 patients were randomized to re- difference 20.53% [95% CI 20.64 to HbA1c of $0.5% (5.5 mmol/mol) after ceive linagliptin (n = 631) or placebo (n = 20.43], 25.8 mmol/mol; P , 0.0001). 52 weeks of treatment (37% linagliptin 630) once daily (Fig. 1). Of these, 1,063 The treatment difference between linaglip- vs. 17% placebo; P , 0.0001). patients (84.3%) completed the trial (543 tin and placebo was maintained for 76 Subgroup analyses of HbA1c changes [86.1%] receiving linagliptin vs. 520 weeks (Fig. 2A). from baseline. Analysis of change in [82.5%] receiving placebo). The main At week 24, the placebo-adjusted HbA1c by prespecified subgroups demon- reasons for discontinuation were AEs decrease in FPG was 20.6 mmol/L strated no significant interaction with

Figure 1dPatient disposition. care.diabetesjournals.org DIABETES CARE 3 Linagliptin as add-on to basal insulin

Table 1dBaseline demographics and clinical characteristics vs. placebo) for rescue medication was 0.575 (95% CI 0.454–0.728; P , 0.0001). Linagliptin Placebo The mean (SD) change in basal insulin dose to week 24 (when dose was to have re- Demographics mained within 10% of baseline) was 0.1 Patients (treated seta) 631 630 (0.2) IU for patients treated with linagliptin Males 329 (52.1) 329 (52.2) and 0.4 (0.2) IU for those treated with pla- Age (years) 59.7 6 9.9 60.4 6 10.0 cebo. From week 24, mean basal insulin Age group dose increased in the linagliptin group ,65 years 431 (68.4) 408 (64.8) to a lesser extent than in the placebo group 65–74 years 165 (26.1) 181 (28.7) (Fig. 2B). The adjusted mean (SD) changes $75 years 35 (5.5) 41 (6.5) from baseline in insulin dose at week 52 Race were 2.6 (0.8) IU for linagliptin and 4.2 (0.8) American Indian/Alaskan Native 4 (0.6) 6 (1.0) IU for placebo (P , 0.003). Asian 80 (12.7) 74 (11.7) Black/African American 41 (6.5) 39 (6.2) Safety and tolerability Hawaiian/Pacific Islander 2 (0.3) 3 (0.5) The overall incidence of patients with $ 1 Caucasian 504 (79.9) 508 (80.6) reported AE was comparable between BMI (kg/m2) 30.8 6 5.4 31.2 6 4.9 treatment groups (linagliptin 78.4%, pla- Renal function (eGFR) according to MDRD cebo 81.4%) (Supplementary Table 2). Normal renal function ($90 mL/min) 277 (43.9) 275 (43.7) AEs were primarily of mild or moderate Mild impairment (60 to ,90 mL/min) 292 (46.3) 283 (44.9) intensity; AEs of severe intensity oc- Moderate impairment (30 to ,60 mL/min) 59 (9.4) 68 (10.8) curred in 8.2% and 8.3% of patients in Severe to end-stage impairment (,30 mL/min) 3 (0.5) 4 (0.6) the linagliptin and placebo groups, re- Clinical characteristics spectively. The most commonly reported Patients (FASb) 618 617 severe AEs in both treatment groups were 6 6 HbA1c (%) 8.31 0.85 8.29 0.85 hypoglycemia, coronary artery disease, 6 6 HbA1c (mmol/mol) 67 9.3 67 9.3 osteoarthritis, cardiac failure, pneumonia, FPG (mmol/L) 8.2 6 2.6 8.4 6 2.6 diarrhea, arthralgia, gastroenteritis, acute Time since diagnosis of diabetes renal failure, and subdural hematoma. AEs #1 year 14 (2.3) 12 (1.9) considered to be drug-related occurred in .1to# 5 years 86 (13.9) 66 (10.7) 18.7% and 22.2% of linagliptin and pla- .5 years 518 (83.8) 539 (87.4) cebo patients, respectively. Drug-related Basal insulin dose (IU/day) 41.5 6 31.9 40.1 6 27.3 hypoglycemia occurred in 83 (15.1%) Concomitant OADs and 95 (13.2%) patients in the linagliptin None 96 (15.5) 102 (16.5) and placebo groups, respectively, and was Metformin only 470 (76.1) 464 (75.2) theonlydrug-relatedAEwithanincidence Pioglitazone only 6 (1.0) 6 (1.0) greater than 2%. There were no imbalances Metformin plus pioglitazone 46 (7.4) 45 (7.3) between treatment groups for other drug- Data are n (%) or mean 6 SD. MDRD, Modification of Diet in Renal Disease Study equation. aAll patients who related AEs. AEs leading to discontinuation were treated with at least one dose of study medication. bAll patients who had a baseline and at least one on- of trial medication occurred in 21 patients treatment HbA1c measurement. (3.3%) in the linagliptin group and 28 patients (4.4%) in the placebo group. There were no clinically relevant changes in vital treatment according to renal function cat- .5 years (20.72% [0.05]). The placebo- signs or laboratory parameters in either egory (P = 0.5784), type of basal insulin adjusted mean change from baseline group, including no between-group imbal- (P = 0.9511), age group (P = 0.1000), con- in HbA1c after 24 weeks was 20.52% ance in shifts in stage of renal impairment. comitant use of OADs (P = 0.64), sex (P = (25.7 mmol/mol) for patients taking The percentage of patients with 0.98), or BMI (P = 0.99), indicating that no OADs at baseline (n =96[15.50%]; investigator-defined hypoglycemia was not none of these factors altered the efficacy of P , 0.0001), 20.67% (27.3 mmol/ different between groups either at week linagliptin. Subgroup variables with signif- mol) for those taking metformin only 24 (linagliptin 22.0%, placebo 23.2%) or icant (P , 0.10) interactions with treat- (n = 470 [76.1%]; P , 0.0001), at the end of treatment (linagliptin 31.4%, 2 2 ment were baseline HbA1c (P = 0.0725), 0.76% ( 8.3 mmol/mol) for those tak- placebo 32.9%). Incidence of severe hypo- geographical region (P = 0.0548), race ing pioglitazone only (n =12[1.0%];P = glycemia was also similar between groups (P = 0.0603), and time since diabetes diag- 0.129), and 20.71% (27.8 mmol/mol) (week 24 linagliptin 0.3%, placebo 0.6%; nosis (P = 0.0017). Placebo-adjusted mean for patients taking both metformin and end of treatment linagliptin 1.7%, placebo (SD) changes in HbA1c were greatest for pioglitazone (n =91[7.4%];P , 0.0001). 1.1%) (Supplementary Table 2). fi patients with a baseline HbA1c of $9.0% Use of rescue therapy and changes in There was no signi cant change from (75 mmol/mol) (20.83% [0.10]), patients background therapy. More patients re- baseline in mean body weight at either who were from Asia (21.00 [0.15]) or of quired rescue medication in the placebo week 24 (linagliptin 20.16 [0.12] kg, Asian race (20.93% [0.14]), and patients group (50.4%) than in the linagliptin placebo 0.12 [0.11] kg) or at week 52 who had been diagnosed with diabetes for group (38.2%); the odds ratio (linagliptin (linagliptin 20.3 [0.19] kg, placebo

4 DIABETES CARE care.diabetesjournals.org Yki-Järvinen and Associates

20.04 [0.18] kg). Adjudicated cardiovascular events occurred in 18 linagliptin patients (2.9%) and 11 placebo patients (1.7%). Cardiovascular deaths occurred in 5 patients (0.8%) in the linagliptin group and 1 (0.2%) in the placebo group. Total mortality, however, was comparable between the arms, with 5 deaths in each group.

CONCLUSIONSdIn patients with inadequate glycemic control despite treatment with OADs, basal insulin therapy is recommended with or without additional OADs (14). This phase III clinical trial dem- onstrated that, in basal-insulin treated patients with type 2 diabetes and inadequate glycemic control (15,16), the addition of linagliptin 5 mg once-daily improved glycemic control without increasing the risk of hypoglycemia or body weight gain. These improvements were not affected by concomitant use of OADs, type of basal insulin, age, or degree of renal impairment. Previous studies have also shown that DPP-4 inhibitors can significantly decrease HbA1c concentration when added to various insulin regimens, with the ex- ception of sitagliptin, without an additional risk of hypoglycemia (8–13). Our study differs from the previous trials in specificallytestingamorehomogenous population of basal insulin–treated patients and including both a period of stable insulin dosing, permitting robust assessment of the efficacy and safety of add-on linagliptin, and an extension period, during which patients were allowed to adjust the insulin according to the investigator criteria. Addition of linagliptin to basal insulin was associated with a low frequency of hypoglycemia, implying that dose reduction of basal insulin to avoid hypoglycemia when coadministering linagliptin may not be necessary. Although the extension period was intended primarily to provide long-term safety data for linagliptin, it also provided additional information on utilization in clinical practice. In the previous trials of other DPP-4 inhibitors, the insulin dose was either kept stable (8,9,13) or titrated (10–12), but not both. During the extension period, the insulin dose was increased more in the placebo group than in the linagliptin group. The magnitude of the increase in insulin dose, d Figure 2 A:MeanHbA1c change from baseline with time out to 76 weeks (FAS, OC). B: Mean however, was much less than recom- FPG change (FAS, OC) and mean change in insulin dose (FAS, original results) from baseline mended in the study protocol (i.e., to achieve with time out to 76 weeks. an FPG target of 6.1 mmol/L). Because no forced titration was requested, this may have resulted in some “titration inertia.” Of note, the proportion of elderly patients care.diabetesjournals.org DIABETES CARE 5 Linagliptin as add-on to basal insulin and patients with renal impairment was S.Pi., S.B., S.T., S.Pa., and H.-J.W. are employ- 5. Drucker DJ, Nauck MA. The incretin high in the study, and it appears likely ees of Boehringer Ingelheim. No other potential system: glucagon-like peptide-1 receptor that investigators may have been more conflicts of interest relevant to this article were agonists and dipeptidyl peptidase-4 in- reported. hibitors in type 2 diabetes. Lancet 2006; reluctant to uptitrate basal insulin be- – cause age and renal impairment both in- H.Y.-J., J.R., S.D.-G., S.Pi., S.B., S.Pa., and 368:1696 1705 H.-J.W participated in the design of the study; 6. Graefe-Mody U, Friedrich C, Port A, et al. crease risk of hypoglycemia. The mean the conduct of the study; the collection, anal- Effect of renal impairment on the pharma- FPG change from baseline at week 52 was ysis, and interpretation of data; and the writing cokinetics of the dipeptidyl peptidase-4 similar between groups, indicating that and revision of the article. S.B. planned and inhibitor linagliptin. Diabetes Obes Metab clinicians were targeting similar FPG goals. performed the statistical analysis of the data. 2011;13:939–946 Nevertheless, significantly lower HbA1c S.T. participated in the collection, analysis, 7. Graefe-Mody U, Rose P, Retlich S, et al. values were obtained with linagliptin. and interpretation of data and in the writing Pharmacokinetics of linagliptin in sub- Patients themselves are also often re- and revision of the article. All the authors were jects with hepatic impairment. Br J Clin luctant to uptitrate insulin because of the fully responsible for all content and editorial Pharmacol 2012;74:75–85 decisions, were involved at all stages of man- 8. Vilsbøll T, Rosenstock J, Yki-Järvinen H, potentially increased risk of hypoglyce- fi mia. This represents a significant consid- uscript development, and have approved the et al. Ef cacy and safety of sitagliptin final version. H.Y.-J. is the guarantor of this eration driving the need for additional when added to insulin therapy in patients work and, as such, had full access to all the with type 2 diabetes. Diabetes Obes Metab treatments such as linagliptin to improve data in the study and takes responsibility for 2010;12:167–177 glycemic control. The low risk of hypogly- the integrity of the data and the accuracy of the 9. Barnett AH, Charbonnel B, Donovan M, cemia with linagliptin could potentially data analysis. Fleming D, Chen R. Effect of saxagliptin as reduce indirect costs, such as hospitali- Data from this study were presented in ab- add-on therapy in patients with poorly zation, and this may counterbalance the stract form at the 72nd Scientific Sessions of controlled type 2 diabetes on insulin alone added direct expense (17,18). the American Diabetes Association, Phila- or insulin combined with metformin. In contrast to other DPP-4 inhibitors, delphia, Pennsylvania, 8–12 June 2012, and at Curr Med Res Opin 2012;28:513–523 linagliptin has a primarily nonrenal route the 48th Annual Meeting of the European 10. Rosenstock J, Rendell MS, Gross JL, Fleck of elimination and therefore does not re- Association for the Study of Diabetes, Berlin, PR, Wilson CA, Mekki Q. Alogliptin added Germany, 1–5 October 2012. quire dose adjustment in patients with to insulin therapy in patients with type 2 The authors thank the patients and staff who diabetes reduces HbA(1C) without causing impaired renal function (6,19). Because participated in this study. (A complete list of weight gain or increased hypoglycaemia. many patients taking insulin have im- participating investigators appears in the Sup- Diabetes Obes Metab 2009;11:1145–1152 paired renal function, this is a potential plementary Data online.) Medical writing as- 11. Fonseca V, Baron M, Shao Q, Dejager S. advantage of linagliptin relative to other sistance, supported financially by Boehringer Sustained efficacy and reduced hypogly- DPP-4 inhibitors. A previous study found Ingelheim, was provided by Claire Stevens and cemia during one year of treatment with linagliptin to be well tolerated and effica- Paul Lane of Envision Scientific Solutions dur- vildagliptin added to insulin in patients cious in long-term use in patients with ing the preparation of this manuscript. with type 2 diabetes mellitus. Horm severe renal impairment (20). Metab Res 2008;40:427–430 In conclusion, this study shows that 12. Fonseca V, Schweizer A, Albrecht D, addition of linagliptin to basal insulin References Baron MA, Chang I, Dejager S. Addition of 1. Rosenstock J, Sugimoto D, Strange P, vildagliptin to insulin improves glycaemic improves glycemic control without increas- Stewart JA, Soltes-Rak E, Dailey G. Triple control in type 2 diabetes. Diabetologia ing hypoglycemia or inducing weight gain, therapy in type 2 diabetes: insulin glargine 2007;50:1148–1155 with the additional advantage that the dose or rosiglitazone added to combination 13. Kothny W, Foley J, Kozlovski P, Shao Q, does not need to be altered in the elderly or therapy of sulfonylurea plus metformin in Gallwitz B, Lukashevich V. Improved in those with impaired renal function. insulin-naive patients. Diabetes Care 2006; glycaemic control with vildagliptin added 29:554–559 to insulin, with or without metformin, in 2. Hermansen K, Davies M, Derezinski T, patients with type 2 diabetes mellitus. AcknowledgmentsdThis study was spon- Martinez Ravn G, Clauson P, Home P. A Diabetes Obes Metab 2013;15:252–257 sored by Boehringer Ingelheim. H.Y.-J. has 26-week, randomized, parallel, treat-to- 14. Inzucchi SE, Bergenstal RM, Buse JB, et al.; received research support from Lilly-Amylin target trial comparing insulin detemir with American Diabetes Association (ADA); and Boehringer Ingelheim. H.Y.-J. has also NPH insulin as add-on therapy to oral European Association for the Study of received honoraria for consultancy for Sanofi, glucose-lowering drugs in insulin-naive Diabetes (EASD). Management of hyper- Novartis, Astra-Zeneca, Merck Sharp & Dohme, people with type 2 diabetes. Diabetes Care glycemia in type 2 diabetes: a patient- and Bristol-Myers Squibb and for lectures at 2006;29:1269–1274 centered approach: position statement of meetings sponsored by Sanofi and Merck Sharp 3. Riddle MC, Rosenstock J, Gerich J; In- the American Diabetes Association (ADA) & Dohme. J.R. has served on scientificadvisory sulin Glargine 4002 Study Investigators. and the European Association for the boards and received honoraria or consulting The treat-to-target trial: randomized Study of Diabetes (EASD). Diabetes Care fees from Roche, Sanofi,NovoNordisk,Eli addition of glargine or human NPH in- 2012;35:1364–1379 Lilly, MannKind, GlaxoSmithKline, Takeda, sulin to oral therapy of type 2 diabetic 15. Hoerger TJ, Segel JE, Gregg EW, Saaddine Daiichi Sankyo, Johnson & Johnson, Novartis, patients. Diabetes Care 2003;26:3080– JB. Is glycemic control improving in U.S. Boehringer Ingelheim, and Lexicon. J.R. has also 3086 adults? Diabetes Care 2008;31:81–86 received grants and research support from 4. Yki-Järvinen H, Juurinen L, Alvarsson M, 16. Willey CJ, Andrade SE, Cohen J, Fuller JC, Merck, Pfizer, Sanofi, Novo Nordisk, Roche, et al. Initiate Insulin by Aggressive Titration Gurwitz JH. Polypharmacy with oral an- Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, and Education (INITIATE): a randomized tidiabetic agents: an indicator of poor Takeda, Novartis, AstraZeneca, Amylin, Johnson study to compare initiation of insulin com- glycemic control. Am J Manag Care 2006; & Johnson, Daiichi Sankyo, MannKind, Lexicon, bination therapy in type 2 diabetic patients 12:435–440 and Boehringer Ingelheim. S.D.-G has received individually and in groups. Diabetes Care 17. Bron M, Marynchenko M, Yang H, Yu consulting fees from Boehringer Ingelheim. 2007;30:1364–1369 AP, Wu EQ. Hypoglycemia, treatment

6 DIABETES CARE care.diabetesjournals.org Yki-Järvinen and Associates

discontinuation, and costs in patients of a large US administrative claims database. 20. McGill JB, Sloan L, Newman J, et al. with type 2 diabetes mellitus on oral an- Ann Pharmacother 2012;46:157–168 Long-term efﬁcacy and safety of lina- tidiabetic drugs. Postgrad Med 2012;124: 19. Blech S, Ludwig-Schwellinger E, Gräfe-Mody gliptin in patients with type 2 diabetes 124–132 EU, Withopf B, Wagner K. The metabolism and severe renal impairment: a 1-year, 18. Simeone JC, Quilliam BJ. Predictors of emer- and disposition of the oral dipeptidyl pep- randomized, double-blind, placebo- gency department and outpatient visits for tidase-4 inhibitor, linagliptin, in humans. controlled study. Diabetes Care 2013;36: hypoglycemia in type 2 diabetes: an analysis Drug Metab Dispos 2010;38:667–678 237–244

care.diabetesjournals.org DIABETES CARE 7