Efficacy and Safety of Glimepiride with Or Without Linagliptin Treatment in Patients with HNF1A Diabetes (Maturity-Onset Diabete

Diabetes Care Volume 43, September 2020 2025

Alexander S. Christensen,1,2

ﬁ CARE/EDUCATION/NUTRITION/PSYCHOSOCIAL CLIN Ef cacy and Safety of Glimepiride Soﬁe Hædersdal,1,2 Julie Støy,3 Heidi Storgaard,1,2 Ulla Kampmann,3 With or Without Linagliptin Julie L. Forman,4 Marta Seghieri,1,2,5 Jens J. Holst,6,7 Torben Hansen,7 Treatment in Patients With Filip K. Knop,1,2,7,8 and Tina Vilsbøll1,2,8 HNF1A Diabetes (Maturity-Onset Diabetes of the Young Type 3): A Randomized, Double-Blinded, Placebo-Controlled, Crossover Trial (GLIMLINA) Diabetes Care 2020;43:2025–2033 | https://doi.org/10.2337/dc20-0408

OBJECTIVE 1Steno Diabetes Center Copenhagen, Gentofte Sulfonylureas are first-line treatment of hepatocyte nuclear factor 1-a (HNF1A) Hospital, Hellerup, Denmark diabetes (maturity-onset diabetes of the young type 3), but many patients do not 2Center for Clinical Metabolic Research, Gentofte achieve optimal glycemic control without episodes of hypoglycemia. We inves- Hospital, University of Copenhagen, Hellerup, Denmark tigated the combination of the sulfonylurea glimepiride and the dipeptidyl 3Steno Diabetes Center Aarhus, Aarhus Univer- peptidase 4 inhibitor linagliptin versus glimepiride monotherapy with respect sity Hospital, Aarhus, Denmark to glycemic variability, glycemic control, and risk of hypoglycemia. 4Section of Biostatistics, Department of Public Health, Faculty of Health and Medical Sciences, RESEARCH DESIGN AND METHODS University of Copenhagen, Copenhagen, Denmark 5 n 5 Diabetes Unit, USL Toscana Centro, Florence, In a randomized, double-blinded, crossover trial, patients with HNF1A diabetes ( Italy 2 19; mean 6 SD age 43 6 14 years, BMI 24.8 6 2.8 kg/m , and glycated hemoglobin 6Department of Biomedical Sciences, Faculty of [HbA1c] 7.4 6 0.2% [57.1 6 7.3 mmol/mol]) were randomly assigned to treatment Health and Medical Sciences, University of Co- 1 penhagen, Copenhagen, Denmark with glimepiride linagliptin 5 mg (16 weeks), washout (4 weeks), and 7 1 Novo Nordisk Foundation Center for Basic Met- glimepiride placebo (16 weeks) (or vice versa). Glimepiride was titrated abolic Research, Faculty of Health and Medical targeting a fasting plasma glucose of 4.5–6.0 mmol/L without hypoglycemia. Sciences, University of Copenhagen, Copenha- Treatments were evaluated by continuous glucose monitoring (CGM), HbA ,and gen, Denmark 1c 8 meal test. Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copen- hagen, Copenhagen, Denmark RESULTS 1 1 fi Corresponding author: Tina Vilsbøll, t.vilsboll@ Compared with glimepiride placebo, glimepiride linagliptin did not signi cantly dadlnet.dk improve the primary end point, mean amplitude of glycemic excursions (MAGE) Received 28 February 2020 and accepted 13 June (mean difference 20.7 mmol/L, P 5 0.1540), but displayed significant reductions 2020 in coefficient of variation on CGM (23.6%, P 5 0.0401), HbA1c (20.5%, P 5 0.0048), Clinical trial reg. no. 2017-000204-15, https:// and glimepiride dose (20.7 mg/day, P 5 0.0099). b-cell glucose sensitivity eudract.ema.europa.eu (assessed as C-peptide–to–glucose ratio) during meal test improved with This article contains supplementary material online glimepiride 1 linagliptin. Incidences of hypoglycemia were similar with both at https://doi.org/10.2337/figshare.12515230. treatments. © 2020 by the American Diabetes Association. Readers may use this article as long as the work is CONCLUSIONS properly cited, the use is educational and not for profit, and the work is not altered. More infor- Linagliptin as add-on treatment to glimepiride improved glycemic variability and mation is available at https://www.diabetesjournals control without increasing risk of hypoglycemia in patients with HNF1A diabetes. .org/content/license. 2026 GLIMLINA Trial Diabetes Care Volume 43, September 2020

Hepatocyte nuclear factor 1-a (HNF1A) by the enzyme dipeptidyl peptidase-4 evaluation, patients were randomly as- diabetes is a monogenic diabetes sub- (DPP-4). Inhibitors of DPP-4 increase signed in a 1:1 ratio to the following two type that belongs to the heterogenous endogenous levels of GIP and GLP-1 treatment sequences: 1) glimepiride 1 group of diabetes subtypes collectively and are well-known for their glucose- linagliptin(16weeks),washout(4weeks), known as maturity-onset diabetes of the lowering actions with low risk of hypogly- and glimepiride 1 placebo (16 weeks) or young (MODY). Approximately 1%–2% of cemia in patients with type 2 diabetes 2) glimepiride 1 placebo (16 weeks), all diabetes cases have a monogenic (20,21). In patients with type 2 diabe- washout (4 weeks), and glimepiride 1 cause. HNF1A diabetes (MODY3) is the tes, the glucose-lowering effect of DPP-4 linagliptin (16 weeks) (Supplementary most prevalent MODY subtype, account- inhibition is mainly attributed to the Fig. 1). The primary end point was the ing for ;50% of all MODY cases (1–3). increase of intact GLP-1 levels because difference between treatments in mean Similar to type 1 diabetes, tight glycemic insulinotropic effect of GIP is severely amplitude of glycemic excursions (MAGE) control is critical for reducing microvas- diminished in acute studies (22). How- calculated from 6 days of continuous cular and macrovascular complications in ever, positive GIP-mediated effects may glucose monitoring (CGM) at the end of patients with HNF1A diabetes (4,5). A key occur during prolonged DPP-4 inhibitor each treatment period. Secondary end characteristic of HNF1A diabetes is im- therapy (23). Recent work from our points were differences between treat- paired glucose-stimulated insulin secre- group has shown additive–to–supra- mentsinregardto1)othervariablesof tion.Thisisduetoareducedb-cellglucose additive insulinotropic effects of GIP glycemic variability calculated from CGM uptake and metabolism resulting in low and GLP-1 in patients with HNF1A di- (coefficient of variation [CV], SD, mean intracellular ATP levels (6,7), which, in abetes treated with sulfonylureas (24). glucose, and low blood glucose index), 2) turn, leave KATP channels open, prevent- In this study, we evaluated the com- HbA1c, 3) fasting plasma glucose, 4) body ing depolarization and insulin release (8). bination of the sulfonylurea glimepir- weight, 5) episodes of hypoglycemia By binding to the sulfonylurea receptor, a ide and the DPP-4 inhibitor linagliptin (assessedwithbothCGMandself- subunit of the KATP channel, sulfonyl- versus glimepiride monotherapy with measured blood glucose), 6) quality of life, ureas close these channels, causing depo- respect to glycemic variability and con- and 7) plasma glucose excursions, pan- larization and insulin release (7,9). The trol (glycated hemoglobin [HbA1c]) and creatic hormone responses (C-peptide potent insulinotropic effect of sulfonyl- risk of hypoglycemia in patients with and glucagon), and acetaminophen ab- ureas and the normal-to-high insulin HNF1A diabetes. sorption during a 4-h combined meal sensitivity typical for patients with HNF1A and bicycle test. diabetes (7,10) explain their potent glucose- RESEARCH DESIGN AND METHODS lowering effect in these patients (11,12). Approvals Patient Population Accordingly, the sulfonylurea gliclazide The study received approval by the Potential participants were screened af- resulted in a 5.2-fold greater reduction in Danish Medicines Agency, the Scientific- ter receiving detailed information about fasting plasma glucose in patients with Ethical Committees of the Central and the trial and signing informed consent. HNF1A diabetes as compared with met- Capital Region of Denmark, and the Dan- Eligibility was evaluated according to formin (11). Thus, sulfonylureas are rec- ish Data Protection Agency. The trial was inclusion and exclusion criteria. Inclu- ommended as first-line therapy in initiated September 2017 and concluded sion criteria were 1) diagnosis of HNF1A patients with HNF1A diabetes (13,14). in June 2019. The trial was registered in diabetes caused by a heterozygous mutation However, sulfonylurea treatment has the European Union Clinical Trials Reg- in HNF1A confirmed by Sanger sequenc- limitations: first, sulfonylureas do not ister (EudraCT) (reg. no. 2017-000204- ing of the gene, 2) monotherapy with a providesustainedglycemiccontrolin all 15) prior to initiation of the trial, and a stable dose of glimepiride of ($0.5 mg patients, and additional glucose-lowering published trial protocol is available (25). per day) during 4 weeks or no glucose- treatment is often needed; second, sulfo- The study was conducted in accordance lowering agent, 3)HbA1c $6.5% ($48 nylurea increases risk of hypoglycemia with the Declaration of Helsinki and In- mmol/mol) (patients on glimepiride treat- (15,16); and third, sulfonylurea may lead ternational Council for Harmonisation of ment) or HbA1c $7.0% ($53 mmol/mol) to increased body weight (17). Accordingly, Technical Requirements for Pharmaceut- (patients receiving no glucose-lowering there is a need for studies investigating icals for Human Use guidelines for good agents), 4) age $18 years, 5) capability to efficient glucose-lowering agents as clinical practice (GCP). The study was perform a 30-min light bicycle test at a add-on to sulfonylurea treatment, when monitored by the GCP unit at the Uni- heart rate of 100–120 bpm, and 6) use of patients lack glycemic control, or an alter- versity of Copenhagen and the GCP unit intrauterine contraceptive devices or native treatment approach, when pa- at the Aarhus University Hospital and hormonal contraception (females). Ex- tients do not achieve glycemic targets University Hospital of Aalborg, Aarhus, clusion criteria were 1) use of glucose- due to recurrent hypoglycemia with Denmark. lowering drugs other than glimepiride; sulfonylureas. 2) uremia, end-stage renal disease, or Glucose-dependent insulinotropic pep- Trial Design and End Points estimated glomerular filtration rate ,30 tide (GIP) and glucagon-like peptide 1 This was a randomized, double-blinded, mL/min/1.73 m2 and/or macroalbumin- (GLP-1) are gut-derived incretin hor- placebo-controlled, crossover trial con- uria; 3) liver disease, serum alanine ami- mones that stimulate insulin release in a ducted at two sites: Steno Diabetes Center notransferase, and/or serum AST .2 3 glucose-dependent manner (i.e., only at Copenhagen (Gentofte Hospital) and upper normal serum levels; 4)anemia plasma glucose levels .4–5mmol/L) Steno Diabetes Center Aarhus (Aarhus (males, hemoglobin ,8.0 mmol/L; females, (18,19). Both hormones are degraded University Hospital). After a 1-week baseline ,7.0 mmol/L); 5) history of acute and/or care.diabetesjournals.org Christensen and Associates 2027

chronic pancreatitis; 6) pregnancy, breast- randomization was immediately reintro- (Panodil; GlaxoSmithKline A/S, Copenha- feeding, or intention to become pregnant; duced and linagliptin/placebo was ter- gen, Denmark). Patients arrived in the 7) inability to complete the study; and 8) minated. Participants were encouraged clinical research facility after an over- known allergic reaction to study medication. not to change their lifestyle (i.e., diet, night fast of at least 10 h. Trial medication exercise, and smoking status) during the was given at time 260 min, and the meal Randomization trial. Episodes of hypoglycemia were de- was served at time 0 min. At time 60– Randomization codes were developed fined according to guidelines and are 90 min, patients cycled with a target by a computer-generated random list of reported separately for the maintenance heart rate of 90–110 bpm. Patients were numbers. and the patients were allo- phase and during CGM and combined monitored for hypoglycemic symptoms, catedtotreatmentgroupbyanun- meal and bicycle test. Patient-reported and blood was sampled for measurements blinded employee who was otherwise episodes of hypoglycemia are defined as of plasma glucose, C-peptide, glucagon, uninvolved in the study. A.S.C. enrolled blood glucose ,4.0 mmol/L confirmed and acetaminophen at prespecified the patients, but both patients and A.S.C. by self-measured blood glucose mea- time points (see Supplementary Fig. 3). remained blinded throughout the trial surements on a glucometer (26). A hypo- period and data analyses. glycemic episode during CGM was defined Analytical Procedures as glucose measured #3.9 mmol/L in at For CGM measurements, iPro2 CGM Trial Medication least 20 consecutive minutes as recom- (Medtronic, Northridge, CA) was used, Tablets with linagliptin 5 mg and placebo mended in guidelines (27). Hypoglyce- which has previously been validated (28). had identical appearance and were pro- mia was categorized as level 1 hypoglycemia For calibration of CGM and measure- vided by Boehringer Ingelheim. Tablets (plasma glucose 3.0–3.9 mmol/L), level 2 hy- ments of blood glucose during the trial, of glimepiride 1 mg (with the possibility poglycemia (plasma glucose ,3.0 mmol/L), participants were given Contour XT gluco- of splitting in half) was not blinded and and level 3 hypoglycemia (severe hypogly- meters (Ascencia Diabetes Care, Co- were delivered by the Hospital Phar- cemia with need of third-party assistance penhagen, Denmark). A description of macy of the Capital Region, Herlev, for recovery) (26). analytical procedures regarding the sam- Denmark. Patients treated with glime- ples from the meal test is presented in piride at screening continued the same Clinical Visits the Supplementary Material. dosage during the 1-week baseline eval- Six types of visits were performed: screen- uation and during washout. One treat- ing, randomization, mounting of CGM, Statistical Analysis and Calculations ment-na¨ıve patient started on 0.5 mg telephone visits, clinical visits, and a com- All analyses were performed on the glimepiride after the baseline evaluation. bined meal and bicycle test. Specification intention-to-treat population, who were Each treatment period was divided into a of blood samples taken at the visits has randomized and received at least one drug titration period (weeks 1–4/weeks previously been published (25). A study dose of linagliptin or placebo. Normally 16–20) and a maintenance period (weeks timeline is depicted in Supplementary distributed data were summarized by num- 5–15/weeks 25–35). During the titration Table 1. At all onsite visits, the following ber of observations (n) and mean and SD. period, glimepiride was uptitrated once were registered: adverse events, epi- Data that were not normally distributed weekly (0.5 mg per week) in a treat-to- sodes of hypoglycemia, and compliance. are presented as median and interquar- target manner with a maximum daily CGM devices were placed on the lower tile ranges. Categorical data were sum- dose of 6 mg in both groups (Supplementary left or right quadrant of the abdomen, marizedinnumbersandpercentages. Fig. 1). A total daily dose of glimepiride andpatientswereinstructedtomea- Comparisons between treatment out- $1 mg was administered in two doses: sure plasma glucose four times daily, comes (except body weight) were per- one dose in the morning and one dose in preferably before the three main meals formed in a linear mixed model including the evening. Target fasting plasma glucose and before bedtime. treatment as a fixed effect and with an (average for 5 days) was between 4.5 and unstructured covariance pattern. P val- 6.0 mmol/L (both inclusive) without epi- Meal and Bicycle Test ues were evaluated using Kenward-Roger sodes of symptomatic or biochemical hy- A combined meal and bicycle test was approximation of the degrees of free- poglycemia (plasma glucose #3.9 mmol/L). applied to make our test similar to that of dom. Missing data were implicitly im- During the maintenance period, all medi- two other studies of patients with HNF1A puted by performing likelihood inference cation doses remained unaltered unless diabetes (15,16). The combined meal and in the linear mixed model. To investigate patients had recurrent (two or more) bicycle test provides a greater risk of potential carryover effects, we tested the hypoglycemic episodes, and in this case, hypoglycemia (due to exercise-induced treatment * period interaction, but no the dose of glimepiride was downtitrated reduction of plasma glucose) compared evidence was detected for any variable. adequately. If target fasting plasma glu- with a traditional meal test executed in Totest whether relatedness had an effect cose was not achieved during the drug an inactive semirecumbent state. In ad- on the outcome, family was added to the titration period, the dose of glimepiride dition, it also mimics daily life better, as it model as a random effect; however, this was adjusted at the investigators’ discre- takes into account that most patients are did not alter any of the statistical results. tion. Once-daily linagliptin 5 mg or pla- not totally inactive 4 h following a meal. Body weight changes (D values) were com- cebo was initiated at the start of each The combined 4-h meal and bicycle test pared between the treatments using a treatment period and the dose kept consisted of a liquid mixed meal (525 kcal: paired t test. All tests were carried out stable throughout the study. During 65 g carbohydrate, 20 g fat, and 21 g protein) with a significance level of 5% without washout, the dose of glimepiride at to which was added 1.5 g acetaminophen adjustment for multiple comparisons. 2028 GLIMLINA Trial Diabetes Care Volume 43, September 2020

Adverse events are summarized qualita- 3.4 mg with glimepiride 1 linagliptin with glimepiride 1 linagliptin versus tively. All statistical analyses were per- and glimepiride 1 placebo, respectively glimepiride 1 placebo was also observed formed in SAS studio, version 9.4, and (mean difference 20.7 mg [95% CI 21.2 (Table 2). graphical presentation was in GraphPad to 20.2], P 5 0.0099) (Table 2). Nine of Prism 8. For calculation of CGM-derived 17 patients received a lower glimepiride Glycemic Control variables, the program EasyGV was used dose during the treatment period with A reduction in HbA1c was observed with (29). Glucose variables included MAGE, glimepiride 1 linagliptin compared with glimepiride 1 linagliptin compared with CV, SD, mean, low blood glucose index, glimepiride 1 placebo (Supplementary glimepiride 1 placebo (20.5% [20.9 to and time spent in hypoglycemia, near Table 2). The remaining eight patients 20.2] (25.6 mmol/mol [29.3 to 21.8]), normoglycemia, and hyperglycemia. Treat- received the same dose of glimepiride P 5 0.0048) (Table 2). The proportions ment responses during the meal and regardless of linagliptin/placebo; three of patients achieving an HbA1c #6.5% bicycle test were quantified as area under of these received the maximal glime- (48 mmol/mol) or a decrement $0.5% the curve (AUC) calculated using the piride dose (6.0 mg glimepiride), while (5 mmol/mol) without hypoglycemia trapezoidal rule or baseline-subtracted the remaining patients received submax- (during the maintenance phase) were AUC (bsAUC). For bsAUC, the average of imal doses of glimepiride. 53% (9 of 17) with glimepiride 1 values at times 210, 25, and 0 min was linagliptin and 44% (8 of 18) with defined as baseline. Sample size calcu- Glycemic Variability From CGM glimepiride 1 placebo (Supplementary lation was based on the primary outcome No significant reduction in the primary Table 2). MAGE (difference between glimepiride 1 end point, MAGE, was observed with linagliptin and glimepiride 1 placebo). glimepiride 1 linagliptin compared with Fasting Plasma Glucose The power of our study (1 2 b)wasset glimepiride 1 placebo (20.7 mmol/L Compared with glimepiride 1 placebo, at 0.8, and, thus, the risk of accepting a false [21.9 to 0.4], P 5 0.1540) (Table 2, Fig. glimepiride 1 linagliptin reduced fasting hypothesis is 0.20 (b 5 0.20) with a two- 1A–C, and Supplementary Table 3). Com- plasma glucose (20.7 mmol/L [21.4 sided significance level at 0.05 (a 5 0.05). pared with glimepiride 1 placebo, gli- to 20.0], P 5 0.0492), and both treat- A minimal relevant difference of 2 mmol/L mepiride 1 linagliptin resulted in significant ments reduced fasting plasma glucose betweentreatmentsinMAGEwaschosen reductions in both CV (23.6% [27.0 to compared with baseline (Table 2). with an expected SD of 2.8 mmol/L based 20.2], P 5 0.0401) and SD (20.4 mmol/L on the work by Saisho et al. (30). Thus, a [20.8 to 20.1], P 5 0.0210), while mean Hypoglycemia total of 16 patients were needed, but due plasma glucose did not change signifi- During CGM, 15 vs. 32 episodes of toriskofdropoutandtoincreasepowerof cantly (20.5 mmol/L [21.4 to 0.4], P 5 hypoglycemia were observed with gli- secondary end points, 20 patients were 0.2341). A trend toward a reduction in mepiride 1 linagliptin versus glimepir- included. A detailed power calculation time spent in hyperglycemia and an in- ide 1 placebo, respectively (Table 3 has previously been published (25). crease in time spent in normoglycemia and Supplementary Table 3). The median

RESULTS Patient Disposition and Baseline Table 1—Baseline characteristics Characteristics Participants, n 19 We assessed 224 patients with HNF1A Female sex, n (%) 11 (58) diabetes for eligibility to enter the study Age (years) 43 (14) (Supplementary Fig. 2). A total of 22 pa- Caucasian, n (%) 19 (100) tientswere screened: 1wasexcludeddue Diabetes duration (years) 20 (8–34) to albuminuria, and 1 withdrew consent BMI (kg/m2) 24.8 (22.0–25.7) before randomization. One participant Fasting plasma glucose (mmol/L) 8.9 (2.3) was excluded from data analysis due to HbA (%) 7.4 (0.7) withdrawal of consent after randomiza- 1c HbA (mmol/mol) 57.1 (7.3) tion but before exposure to study drugs. 1c HOMA-IR 0.9 (0.7–1.3) Two other participants dropped out dur- Complications, n ing the study in week 4 and week 13 (due Retinopathy 5 fi to dif culty adhering to study protocol Microalbuminuria 1 and time concerns, respectively), and both Transient ischemic attack 1 patients are included in the analysis. Par- Comorbidities, n ticipant baseline characteristics are pre- Hypertension 3 sented in Table 1 and the specific HNF1A Hypercholesterolemia 9 mutations in Supplementary Table 2. Over- Multiple sclerosis 1 all compliance was good: 97% with both Treatment, n glimepiride and linagliptin/placebo. Glimepiride 18 Diet 1 Glimepiride Dose Data are mean (SD) or median (interquartile range) unless otherwise indicated. Diabetes duration fi $ $ The mean glimepiride dose increased is from manifest diabetes ( rst measurement of HbA1c 6.5% [ 48 mmol/mol]). HOMA-IR, HOMA of insulin resistance. from 1.9 mg at baseline to 2.7 mg and Table 2—Primary and secondary outcomes Glimepiride 1 placebo Glimepiride 1 linagliptin Glimepiride 1 linagliptin Baseline vs. baseline vs. baseline vs. glimepiride 1 placebo CGM Christensen and Associates 2029 MAGE (mmol/L) 5.5 (4.6–6.4) 0.3 (20.9 to 1.5), P 5 0.5616 20.4 (21.6 to 0.7), P 5 0.4286 20.7 (21.9 to 0.3), P 5 0.1540 CV (%) 27.2 (24.0–30.5) 4.2 (23.0 to 4.2), P 5 0.0438 0.6 (23.0 to 4.2), P 5 0.7364 23.6 (27.0 to 20.2), P 5 0.0401 SD (mmol/L) 2.6 (2.3–3.0) 0.1 (20.3 to 0.6), P 5 0.5192 20.4 (20.7 to 0.1), P 5 0.1000 20.4 (20.8 to 20.1), P 5 0.0210 Mean (mmol/L) 9.6 (8.9–10.4) 20.9 (21.8 to 0.0), P 5 0.0548 21.4 (22.2 to 20.6), P 5 0.0012 20.5 (21.4 to 0.4), P 5 0.2341 Low blood glucose index 0.9 (0.4–1.5) 0.7 (0.1–1.4), P 5 0.0364 0.4 (20.3 to 1.2), P 5 0.2458 20.3 (21.0 to 0.4), P 5 0.3740 % time in hyperglycemia ($10.0 mmol/L) 38.9 (29.2–48.6) 29.8 (221.2 to 1.7), P 5 0.0894 215.7 (226.0 to 25.5), P 5 0.0047 26.0 (216.4 to 4.5), P 5 0.2470 % time in near normoglycemia (4.0–9.9 mmol/L) 60.7 (50.9–70.4) 8.5 (22.9 to 20.0), P 5 0.1358 15.3 (4.7–25.8), P 5 0.0068 6.7 (23.9 to 17.4), P 5 0.1990 Biochemistry and other HbA1c (%) 7.4 (7.0–7.7) 20.4 (20.8 to 0.1), P 5 0.1165 20.9 (21.3 to 20.5), P 5 0.0001 20.5 (20.9 to 20.2), P 5 0.0048 HbA1c (mmol/mol) 57.1 (53.6–60.5) 24.0 (29.0 to 1.0), P 5 0.1116 29.6 (213.7 to 25.4), P 5 0.0001 25.6 (29.3 to 21.8), P 5 0.0055 Fasting plasma glucose (mmol/L) 9.0 (7.9–10.0) 21.4 (22.5 to 20.3), P 5 0.0182 22.1 (23.2 to 21.0), P 5 0.0008 20.7 (21.4 to 20.0), P 5 0.0492 Mental component score (SF-36) 53.6 (49.5–57.8) 23.2 (26.7 to 0.4), P 5 0.0765 0.6 (22.5 to 3.8), P 5 0.6751 3.8 (21.7 to 9.3), P 5 0.1655 Physical component score (SF-36) 55.2 (52.2–58.2) 0.8 (20.8 to 2.4), P 5 0.2891 2.1 (0.6–3.5), P 5 0.0087 1.2 (20.0 to 2.5), P 5 0.0569 Glimepiride dose (mg) 1.9 (1.3–2.4) 1.5 (0.9–2.1), P , 0.0001 0.8 (0.2–1.5), P 5 0.0166 20.7 (21.2 to 20.2), P 5 0.0099 Meal and bicycle test Plasma glucose AUC (mol/L 3 min) 2.9 (2.5–3.4) 20.3 (20.5 to 20.0), P 5 0.0463 20.6 (20.9 to 20.3), P 5 0.008 20.3 (20.6 to 20.0), P 5 0.0466 bsAUC (mol/L 3 min) 0.9 (0.6–1.2) 20.1 (20.3 to 0.2), P 5 0.6066 20.2 (20.4 to 0.0), P 5 0.0975 20.1 (20.4 to 0.1), P 5 0.2418 Excursion (mmol/L) 7.8 (6.3–9.3) 20.1 (21.3 to 1.1), P 5 0.8478 20.6 (21.8 to 0.7), P 5 0.3396 20.5 (21.9 to 1.0), P 5 0.4976 Peak (mmol/L) 15.3 (13.2–17.5) 20.9 (22.2 to 0.4), P 5 0.1482 22.3 (23.8 to 20.8), P 5 0.0042 21.4 (23.2 to 0.3), P 5 0.1090 C-peptide AUC (nmol/L 3 min) 226 (174–278) 27(223 to 93), P 5 0.3847 2 (215 to 19), P 5 0.7813 9 (23.5 to 22), P 5 0.1453 bsAUC (nmol/L 3 min) 142 (107–176) 221 (239 to 24), P 5 0.0174 25(225 to 15), P 5 0.6073 17 (23 to 36), P 5 0.0940 C-peptide–to–glucose ratio AUC (pmol/L 3 mmol21 3 1023 3 min) 20 (14–26) 2 (22 to 6), P 5 0.2588 7 (3–12), P 5 0.0018 5 (2–8), P 5 0.0008 bsAUC (pmol/L 3 mmol21 3 1023 3 min) 9 (5–13) 1 (23 to 2), P 5 0.5979 4 (22 to 9), P 5 0.1603 4 (1–8), P 5 0.0261 Glucagon Fasting levels (pmol/L) 7.4 (5.6–9.1) 0.1 (21.3 to 1.6), P 5 0.8473 20.7 (22.4 to 1.0), P 5 0.4108 20.8 (22.3 to 0.7), P 5 0.2661 AUC (nmol/L 3 min) 3.1 (2.6–3.7) 20.0 (20.4 to 0.4), P 5 0.8962 20.2 (20.5 to 0.1), P 5 0.1535 20.2 (20.5 to 0.1), P 5 0.2458 bsAUC (nmol/L 3 min) 1.3 (0.9–1.7) 20.0 (20.4 to 0.4), P 5 0.9280 0.0 (20.4 to 0.4), P 5 0.9731 0.0 (20.2 to 0.3), P 5 0.8538 Acetaminophen AUC (mmol/L 3 min) 15.4 (14.1–16.8) 0.0 (20.7 to 0.8), P 5 0.9159 20.6 (21.4 to 0.1), P 5 0.0879 20.7 (21.4 to 20.0), P 5 0.0479 Time to peak (min) 122 (110–133) 23(218 to 13), P 5 0.7337 11 (22 to 25), P 5 0.0936 14 (2–26), P 5 0.0238 Peak (mmol/L) 0.10 (0.09–0.11) 0.01 (20.05 to 0.02), P 5 0.2448 0.00 (20.01 to 0.01), P 5 0.8526 20.01 (20.01 to 20.00), P 5 0.0439 Data are mean (95% CI). SF-36, 36-Item Short Form Health Survey, version 2. care.diabetesjournals.org 2030 GLIMLINA Trial Diabetes Care Volume 43, September 2020

the meal and bicycle test in both treatment arms.

Body Weight There was a signiﬁcant increase in body weight with glimepiride 1 placebo (1.2 kg [0.4–2.0], P 5 0.0078). Glimepiride 1 linagliptin was body weight neutral (0.2 kg [20.7 to 1.0], P 5 0.7025). The mean difference in body weight changes between treatment arms was 21.0 kg (22.3 to 0.2), P 5 0.0961.

Meal and Bicycle Test Time courses for plasma/serum glucose, glucagon, C-peptide, C-peptide–to–glucose ratio, and acetaminophen are presented in Fig. 1. Glimepiride 1 linagliptin resulted in a greater reduction in plasma glucose AUC compared with glimepiride 1 placebo, and both treatment regi- mens reduced AUC for plasma glucose compared with baseline (Table 2). Like- wise, the b-cell sensitivity for glucose assessed by AUC for C-peptide–to–glucose ratio was signiﬁcantly greater with glimepiride 1 linagliptin compared with glimepiride 1 placebo (Fig. 1 and Table 2). An increase in postprandial glucagon levels from time 0 min was observed regardless of treatment (Fig. 1). Gastric emptying assessed by acetaminophen absorption was slower with glimepiride 1 linagliptin compared with glimepiride 1 placeboassessedbyAUC,timetopeak,and peak concentration (Fig. 1 and Table 2).

Quality of Life No significant changes were observed in mental component scores or physical component scores between treatment Figure 1—CGM and meal and bicycle test. A: Mean 6 SD values from CGM during 24 h at baseline arms (Table 2). and at the end of two 16-week treatment periods of glimepiride 1 linagliptin and glimepiride 1 placebo in 19 patients with HNF1A diabetes. B: Differences in means and 95% CI in glycemic Adverse Events variability calculated from CGM data between glimepiride 1 linagliptin and glimepiride 1 placebo. C: Mean percent time spent in different plasma glucose (PG) ranges, calculated from CGM. D–H: No serious adverse events occurred dur- Combined meal and bicycle test concentration versus time for plasma/serum glucose (D), ing the trial. Hypoglycemia was the most C-peptide (E), C-peptide–to–glucose ratio (F), glucagon (G), and acetaminophen (H). Graphics data frequently reported adverse event. Non- are mean 6 SEM if not otherwise indicated. Orange circle, baseline; red square, glimepiride 1 hypoglycemic adverse event counts and 1 , placebo; blue triangle, glimepiride linagliptin. *P 0.05. types were similar with both treatments (Supplementary Table 4). time spent in a state of hypoglycemia was (Table 3 and Supplementary Table 2). in general short: 0.0% (baseline), 0.1% Dose reductions of glimepiride due to CONCLUSIONS (glimepiride 1 linagliptin), and 0.3% recurrent hypoglycemia occurred more This randomized, double-blinded, cross- (glimepiride 1 placebo), respectively (Ta- frequently during the maintenance pe- over study is the first to evaluate add-on ble 3). In contrast, when hypoglycemic riod in the glimepiride 1 linagliptin arm treatment to sulfonylurea in patients with events were assessed from self-reported (two reductions in one patient and one HNF1A diabetes. We find that addition of and self-measured blood glucose during reduction in two patients) compared the DPP-4 inhibitor linagliptin to the sulfo- the maintenance period, a total of 16 and with the glimepiride 1 placebo arm nylurea glimepiride does not improve MAGE 10 episodes of hypoglycemia were ob- (one reduction in one patient). No severe but improves CV, SD, HbA1c,andb-cell served with glimepiride 1 linagliptin ver- hypoglycemia was reported. Two events glucose sensitivity without increasing sus glimepiride 1 placebo, respectively of hypoglycemia were observed during the risk of hypoglycemia. care.diabetesjournals.org Christensen and Associates 2031

Table 3—Hypoglycemia Baseline Glimepiride 1 placebo Glimepiride 1 linagliptin (n 5 19) (n 5 18) (n 5 17) Hypoglycemia during CGM Total episodes 8 32 15 Level 1 episodes 6 30 14 Level 2 episodes 2 2 1 Patients without episode 17 6 12 Percent time spend in hypoglycemia during CGM Total 0.0 (0.0–0.0) 0.3 (0–3.4) 0.1 (0–3.4) Level 1 0.0 (0.0–0.0) 0.3 (0.0–3.0) 0.1 (0.0–0.7) Level 2 0.0 (0.0–0.0) 0.0 (0.0–0.0) 0.0 (0.0–0.0) Patient-reported hypoglycemia in maintenance phase (weeks 5–15 and 24–35) Total episodes N/A 10 16 Level 1 episodes 10 15 Level 2 episodes 01 Patients without episode 11 10 Hypoglycemia during meal and bicycle test Level 1 episodes 1 2 2 Level 2 episodes 0 0 0 Data are n (counts) or median (interquartile range). A hypoglycemic episode verified by CGM is defined as glucose measurements #3.9 mmol/L in at least 20 consecutive minutes. Patient-reported episodes of hypoglycemia are defined as blood glucose ,4.0 mmol/L confirmed by self-measured blood glucose measurements on a glucometer. Level 1 hypoglycemia is a glucose value of 3.0–3.9 mmol/L and level 2 hypoglycemia is defined as a glucose value ,3.0 mmol/L. Both level 1 and level 2 hypoglycemia are per definition nonsevere hypoglycemia (no need of third-party assistance for recovery of hypoglycemia).

A crossover design was chosen to consensus on the topic. During the con- effect (34) and that sulfonylurea-induced ensure enough power given the low duct of our study, a consensus report on insulin secretion can be potentiated by prevalence of HNF1A diabetes and lessen how to report results on glycemic vari- addition of exogenous GIP and GLP-1 recruitment struggles. For reduction of a ability obtained from CGM was pub- infusions, respectively (24). In line with potential carryover effect, a 4-week lished. This strongly recommends CV these observations, we here demon- washout period was introduced between and SD as primary and secondary key strate an improved b-cell sensitivity to treatments. A limitation to our study is metrics, respectively, of glucose vari- glucose (as evaluated from the increase that four patients were related (mother ability (27). Considering this publication, in AUC for C-peptide–to–glucose ratio and son in two families); however, our it is fair to say that linagliptin added to during the combined meal and bicycle statistical analyses indicate no effect of glimepiride treatment in the current test) with glimepiride 1 linagliptin com- family. Another limitation is that our study improved glycemic variability as pared with glimepiride 1 placebo. This study investigated linagliptin given as an assessed by CV and SD. finding is in line with the effect of DPP-4 add-on therapy before the therapeutic In interpretation of the results from inhibitors in patients with type 2 diabetes potential of a dose escalation of glime- our study, it is key to bear in mind that we (35,36). Clinically, this may translate into piride had been fully explored in all used a treat-to-target approach. This improvements of long-term glycemic con- patients. Based on number of partic- resulted in a dose reduction of glime- trol and low glycemic variability. Given the ipants and time span, our study is the piride in 9 of 17 patients (a significant previous study showing insulinotropic largest interventional study performed mean dose reduction of 0.7 mg) during actions of both GIP and GLP-1, it seems in patients with HNF1A diabetes; nev- the glimepiride 1 linagliptin period com- likely that the elevated levels of both ertheless, sample size is small and may pared with the glimepiride 1 placebo hormones due to DPP-4 inhibition con- reduce generalizability of our results. period. Given the high efficacy of sulfo- tribute to the insulinotropic effect ob- Of note, we only included adults ($18 nylureas in patients with HNF1A diabe- served in the current study. years), and our results may not apply to tes, demonstrated in other clinical studies We observed a discrepancy in the pediatric and adolescent populations (11,15), we believe that this is a clinically number of episodes of hypoglycemia with HNF1A diabetes. Three patients relevant dose reduction. measured with CGM (glimepiride 1 li- dropped out during the trial, but our The rationale for investigating DPP-4 nagliptin vs. glimepiride 1 placebo: 15 statistical approach with linear mixed inhibition in the current study originates vs. 32 events, respectively) and self- models implicitly imputes missing data from clinical experience with the inhib- reported events during the maintenance by performing likelihood inference. itors in patients with HNF1A diabetes and phase (glimepiride 1 linagliptin vs. gli- We chose MAGE as the primary end their well-known mode of action on the mepiride 1 placebo: 16 vs. 10 events). point, as it previously has been seen as a b-cells (32,33). Furthermore, our group Our interpretation of these findings is gold standard measurement of glycemic has previously shown that patients with that the combination therapy was indeed variability (31), though without a clear HNF1Adiabeteshaveanimpairedincretin efficacious and that an increased number 2032 GLIMLINA Trial Diabetes Care Volume 43, September 2020

of HNF1A patients needed dose reduc- has given paid lectures for Novo Nordisk and 8. Wollheim CB. Beta-cell mitochondria in the tions of glimepiride (due to recurrent Merck Sharp & Dohme (MSD). F.K.K. has served regulation of insulin secretion: a new culprit fi – hypoglycemia during the maintenance on scienti c advisory panels for, been part of in type II diabetes. Diabetologia 2000;43:265 speakers bureaus for, served as a consultant to, 277 phase) when treated with glimepiride 1 and/or received research support from Amgen, 9. Hansen T, Eiberg H, Rouard M, et al. Novel linagliptinversusglimepiride1 placebo. AstraZeneca, Bayer, Boehringer Ingelheim, Car- MODY3 mutations in the hepatocyte nuclear A dose titration targeted on fasting blood mot Therapeutics, Eli Lilly, Gubra, MedImmune, factor-1a gene: evidence for a hyperexcitability glucose levels of 4.5–6.0 mmol/L and MSD/Merck, Mundipharma, Norgine, Novo Nor- of pancreatic b-cells tointravenoussecretagogues disk, Sanofi, and Zealand Pharma. T.V. has served in a glucose-tolerant carrier of a P447L mutation. absence of hypoglycemia over a time on scientific advisory panels for, served on speakers Diabetes 1997;46:726–730 course of 4 weeks may have been too bureaus for, served as a consultant to, and/or 10. Tripathy D, Carlsson A-L,˚ Lehto M, Isomaa B, short to fully evaluate the risk of hypo- received research support from Amgen, Astra- Tuomi T, Groop L. Insulin secretion and insulin glycemia. This is probably why some pa- Zeneca, Boehringer Ingelheim, Bristol-Myers sensitivity in diabetic subgroups: studies in the tients had recurrent hypoglycemia and Squibb, Eli Lilly, MSD, Mundipharma, Novo Nor- prediabetic and diabetic state. Diabetologia 2000; disk, Sanofi, and Sun Pharma. No other potential 43:1476–1483 dose reductions of glimepiride during the conflicts of interest relevant to this article were 11. Pearson ER, Starkey BJ, Powell RJ, Gribble maintenance phase. Given the differen- reported. FM, Clark PM, Hattersley AT. Genetic cause of ces in observed hypoglycemia between Author Contributions. A.S.C., S.H., H.S., T.H., hyperglycaemia and response to treatment in the methods used, we believe it is fair to F.K.K., and T.V. designed the study and wrote the diabetes. Lancet 2003;362:1275–1281 conclude that the risk of hypoglycemia is study protocol. A.S.C., J.S., and M.S. performed 12. Pearson ER, Liddell WG, Shepherd M, Corrall the study. J.J.H. measured glucagon. J.L.F. as- similar with both treatments. RJ, Hattersley AT. Sensitivity to sulphonylureas in sisted with the statistical analyses. A.S.C. did the patients with hepatocyte nuclear factor-1a gene We found that linagliptin added to data analysis and wrote the first manuscript mutations: evidence for pharmacogenetics in di- glimepiride treatment exerted small but draft. All authors critically edited the manuscript abetes. Diabet Med 2000;17:543–545 statistically significant reductions in gas- andapprovedthefinalversion.A.S.C.andT.V.are 13. ISPAD Clinical Practice Consensus Guidelines tric emptying (evaluated by the acet- the guarantors of this work and, as such, had full 2018 [Internet], 2018. Available from https:// access to all the data in the study and take aminophen absorption test) (Table 2). www.ispad.org/page/ISPADGuidelines2018. Ac- responsibility for the integrity of the data and cessed 13 December 2019 Studies in patients with type 2 diabetes the accuracy of the data analysis. 14. American Diabetes Association. 9. Pharma- have not shown any effect of DPP-4 Prior Presentation. Parts of this study were cologic approaches to glycemic treatment: Stand- inhibition on gastric emptying as as- presented in abstract form at the 80th Scientific ards of Medical Care in Diabetesd2019.Diabetes Sessions of the American Diabetes Association, Care 2019;42(Suppl. 1):S90–S102 sessed by radioactively labeled meals 12–16 June 2020. (37) or the acetaminophen absorption 15. Østoft SH, Bagger JI, Hansen T, et al. Glucose- lowering effects and low risk of hypoglycemia in test (35), and the present finding is likely References patients with maturity-onset diabetes of the young of no or little clinical relevance and may when treated with a GLP-1 receptor agonist: 1. Pihoker C, Gilliam LK, Ellard S, et al.; SEARCH fi fi a double-blind, randomized, crossover trial. Di- represent a chance nding. Our nding for Diabetes in Youth Study Group. Prevalence, abetes Care 2014;37:1797–1805 could also be attributable to the timing of characteristics and clinical diagnosis of maturity blood samples with infrequent blood onset diabetes of the young due to mutations in 16. Tuomi T, Honkanen EH, Isomaa B, Sarelin L, HNF1A,HNF4A,and glucokinase: resultsfrom the Groop LC. Improved prandial glucose control sampling near time to peak, which may with lower risk of hypoglycemia with nateglinide reduce the precision of the acetamin- SEARCH for Diabetes in Youth. J Clin Endocrinol Metab 2013;98:4055–4062 than with glibenclamide in patients with maturity- ophen absorption test. 2. Shields BM, Hicks S, Shepherd MH, Colclough onset diabetes of the young type 3. Diabetes Care – In conclusions, the DPP-4 inhibitor li- K, Hattersley AT, Ellard S. Maturity-onset diabe- 2006;29:189 194 nagliptin added to treatment with the tes of the young (MODY): how many cases are we 17. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional sulfonylurea glimepiride in patients with missing? Diabetologia 2010;53:2504–2508 3. Colclough K, Bellanne-Chantelot C, Saint- treatment and risk of complications in patients HNF1A diabetes reduced glycemic vari- Martin C, Flanagan SE, Ellard S. Mutations in with type 2 diabetes (UKPDS 33). UK Prospective ability (as evaluated from CV and SD but the genes encoding the transcription factors Diabetes Study (UKPDS) Group [published cor- not MAGE) and HbA without increasing hepatocyte nuclear factor 1 alpha and 4 alpha rection appears in Lancet 1999;354:602]. Lancet 1c – the risk of hypoglycemia or other adverse in maturity-onset diabetes of the young and 1998;352:837 853 18. Christensen MB, Calanna S, Holst JJ, Vilsbøll events. hyperinsulinemic hypoglycemia. Hum Mutat 2013;34:669–685 T, Knop FK. Glucose-dependent insulinotropic 4. Isomaa B, Henricsson M, Lehto M, et al. polypeptide: blood glucose stabilizing effects in Chronic diabetic complications in patients with patients with type 2 diabetes. J Clin Endocrinol – Acknowledgments. The authors thank the MODY3 diabetes. Diabetologia 1998;41:467– Metab 2014;99:E418 E426 study participants for commitment and loyalty. 473 19. Weir GC, Mojsov S, Hendrick GK, Habener JF. The authors thank Sisse Marie Schmidt, Inass Al 5. Steele AM, Shields BM, Shepherd M, Ellard S, Glucagonlike peptide I (7-37) actions on endo- Nachar, Kirsten Abelin (all from Steno Diabetes Hattersley AT, Pearson ER. Increased all-cause crine pancreas. Diabetes 1989;38:338–342 Center Copenhagen, Gentofte Hospital, Hellerup, and cardiovascular mortality in monogenic di- 20. Mari A, Sallas WM, He YL, et al. Vildagliptin, a Denmark), and Eva Schriver (Steno Diabetes Cen- abetes as a result of mutations in the HNF1A dipeptidyl peptidase-IV inhibitor, improves model- ter Aarhus, Aarhus University Hospital, Aarhus, gene. Diabet Med 2010;27:157–161 assessed b-cell function in patients with type 2 Denmark) for laboratory assistance. The authors 6. Stride A, Vaxillaire M, Tuomi T, et al. The diabetes. J Clin Endocrinol Metab 2005;90:4888– thank coworkers in their departments for valuable genetic abnormality in the beta cell determines 4894 advice and support. the response to an oral glucose load. Diabeto- 21. McGill JB. Linagliptin for type 2 diabetes Duality of Interest. The study was funded by an logia 2002;45:427–435 mellitus: a review of the pivotal clinical trials. unrestricted grant from Boehringer Ingelheim, 7. Pearson ER, Velho G, Clark P, et al. beta-cell Ther Adv Endocrinol Metab 2012;3:113–124 Copenhagen, Denmark. H.S. has served on ad- genes and diabetes: quantitative and qualitative 22. Mentis N, Vardarli I, Kothe¨ LD, et al. GIP does visory panels for Novo Nordisk and has given paid differences in the pathophysiology of hepatic notpotentiatethe antidiabetic effectsof GLP-1in lectures for Novo Nordisk. J.J.H. has served on nuclear factor-1alpha and glucokinase mutations. hyperglycemic patients with type 2 diabetes. scientific advisory panels for Novo Nordisk and Diabetes 2001;50(Suppl. 1):S101–S107 Diabetes 2011;60:1270–1276 care.diabetesjournals.org Christensen and Associates 2033

23. Aaboe K, Akram S, Deacon CF, Holst JJ, glucose monitoring. Diabetes Care 2017;40: young patients–report of two cases. Diabetes Madsbad S, Krarup T. Restoration of the insu- 1631–1640 Technol Ther 2010;12:313–316 linotropic effect of glucose-dependent insulino- 28. Liebl A, Henrichs HR, Heinemann L, 33. Misra S, Owen KR. Genetics of monogenic tropic polypeptide contributes to the antidiabetic FreckmannG,BiermannE,ThomasA;Continuous diabetes: present clinical challenges. Curr Diab effect of dipeptidyl peptidase-4 inhibitors. Diabe- Glucose Monitoring Working Group of the Work- Rep 2018;18:141 tes Obes Metab 2015;17:74–81 ing Group Diabetes Technology of the German 34. Østoft SH, Bagger JI, Hansen T, et al. Incretin 24. Christensen AS, Hædersdal S, Storgaard H, Diabetes Association. Continuous glucose mon- effect and glucagon responses to oral and in- et al. GIP and GLP-1 potentiate sulfonylurea- itoring: evidence and consensus statement for travenous glucose in patients with maturity- induced insulin secretion in hepatocyte nuclear clinical use. J Diabetes Sci Technol 2013;7:500– onset diabetes of the young–type 2 and type 3. factor 1-a mutation carriers. Diabetes. 9 June 519 Diabetes 2014;63:2838–2844 2020 [Epub ahead of print]. DOI: 10.2337/db20- 29. Hill NR, Oliver NS, Choudhary P, Levy JC, 35. DeFronzo RA, Okerson T, Viswanathan P, 0074 Hindmarsh P, Matthews DR. Normal reference Guan X, Holcombe JH, MacConell L. Effects of 25. Sidelmann Christensen A, Storgaard H, range for mean tissue glucose and glycemic exenatide versus sitagliptin on postprandial glu- Hædersdal S, Hansen T, Krag Knop F, Vilsbøll T. variability derived from continuous glucose mon- cose, insulin and glucagon secretion, gastric Glimepiride monotherapy versus combination of itoring for subjects without diabetes in different emptying, and caloric intake: a randomized, glimepiride and linagliptin therapy in patients ethnic groups. Diabetes Technol Ther 2011;13: cross-over study. Curr Med Res Opin 2008;24: with HNF1A-diabetes: a protocol for a rando- 921–928 2943–2952 mised, double-blinded, placebo-controlled trial. 30. Saisho Y, Tanaka C, Tanaka K, et al. Relation- 36. Ahren´ B, Landin-Olsson M, Jansson P-A, BMJ Open 2018;8:e022517 ships among different glycemic variability indices Svensson M, Holmes D, Schweizer A. Inhibition 26. International Hypoglycaemia Study Group. obtained by continuous glucose monitoring. of dipeptidyl peptidase-4 reduces glycemia, sus- Glucose concentrations of less than 3.0 mmol/L Prim Care Diabetes 2015;9:290–296 tains insulin levels, and reduces glucagon levels (54 mg/dL) should be reported in clinical trials: 31. Baghurst PA. Calculating the mean ampli- in type 2 diabetes. J Clin Endocrinol Metab 2004; a joint position statement of the American Di- tude of glycemic excursion from continuous glu- 89:2078–2084 abetes Association and the European Association cose monitoring data: an automated algorithm. 37. Vella A, Bock G, Giesler PD, et al. Effects of for the Study of Diabetes. Diabetes Care 2017;40: Diabetes Technol Ther 2011;13:296–302 dipeptidyl peptidase-4 inhibition on gastrointes- 155–157 32. Katra B, Klupa T, Skupien J, et al. Dipeptidyl tinal function, meal appearance, and glucose 27. Danne T, Nimri R, Battelino T, et al. In- peptidase-IV inhibitors are efﬁcient adjunct ther- metabolism in type 2 diabetes. Diabetes 2007; ternational consensus on use of continuous apy in HNF1A maturity-onset diabetes of the 56:1475–1480