Diabetes Ther (2019) 10:1893–1907 https://doi.org/10.1007/s13300-019-0671-x ORIGINAL RESEARCH Treatment of Type 2 Diabetes by Patient Profile in the Clinical Practice of Endocrinology in Spain: Delphi Study Results from the Think Twice Program Carlos Morillas . Javier Escalada . Rafael Palomares . Diego Bellido . Fernando Gómez-Peralta . Antonio Pérez Received: May 16, 2019 / Published online: July 29, 2019 Ó The Author(s) 2019 ABSTRACT endocrinologists, a Delphi questionnaire of 55 statements was developed and responded to by Introduction: The aim of this Delphi study is to a national panel (n = 101). unveil the management of patients with type 2 Results: A consensus was reached for 30 of the diabetes (T2D) and different levels of complex- 55 statements. Regarding overweight patients ity in the clinical practice in Spain. inadequately controlled with metformin, treat- Methods: Based on the common management ment with a sodium-glucose transport protein 2 practices of T2D profiles reported by Spanish inhibitor (SGLT2-I) is preferred over treatment with a dipeptidyl peptidase-4 inhibitor (DPP4-I). If the patient is already being treated with a Enhanced Digital Features To view enhanced digital features for this article go to https://doi.org/10.6084/ DPP4-I, an SGLT2-I is added on to the treatment m9.figshare.8869955. regimen rather than replacing the DPP4-I. Conversely, if the treatment regimen includes a Electronic Supplementary Material The online version of this article (https://doi.org/10.1007/s13300- sulfonylurea, it is usually replaced by other 019-0671-x) contains supplementary material, which is antihyperglycemic agents. Current treatment available to authorized users. trends in uncontrolled obese patients include the addition of an SGLT2-I or a glucagon-like C. Morillas (&) peptide-1 receptor agonist (GLP1-RA) to back- Hospital Universitario Doctor Peset, Valencia, Spain e-mail: cmorillas@telefonica.net ground therapy. When the glycated hemoglo- bin target is not reached, triple therapy with J. Escalada metformin ? GLP1-RA ? SGLT2-I is initiated. Clı́nica Universidad de Navarra, Pamplona, Spain Although SGLT2-Is are the treatment of choice R. Palomares in patients with T2D and heart failure or Hospital Universitario Reina Sofı́a, Córdoba, Spain uncontrolled hypertension, no consensus was reached regarding the preferential use of SGLT2- D. Bellido Hospital Arquitecto Marcide, Complejo Hospitalario Is or GLP1-RAs in patients with established Universitario de Ferrol [CHUF], El Ferrol, Spain cardiovascular disease. Conclusion: Consensus has been reached for a F. Gómez-Peralta Hospital General de Segovia, c/Luis Erik Claverı́a variety of statements regarding the manage- neurólogo s/n, Segovia, Spain ment of several T2D profiles. Achieving a more homogeneous management of complex A. Pérez Hospital de la Santa Creu i Sant Pau, CIBERDEM, patients with T2D may require further evidence c/de Sant Quintı́, Barcelona, Spain 1894 Diabetes Ther (2019) 10:1893–1907 and a better understanding of the key drivers for with overlapping comorbidities and circum- treatment choice. stances, and the availability of a wide range of Funding: Logistic support was provided by antihyperglycemic agents with varying efficacy ESTEVE Pharmaceuticals S.A Spain. and side effects make the choice of the appro- priate therapeutic program a difficult issue. Keywords: Clinical practice; Complex patient; The Delphi study reported here was designed Delphi questionnaire; Endocrinology; Type 2 to gather together the opinions of a represen- diabetes tative panel of Spanish endocrinologists spe- cialized in diabetes management on the available treatment options and possible ther- INTRODUCTION apy combinations for different T2D patient profiles with different levels of complexity and Current estimates show that there are 425 mil- to identify common practices that may be used lion adults worldwide with type 2 diabetes as guidance for other physicians. (T2D), of whom 3.6 million are in Spain [1]. Although T2D has a genetic component, multiple risk factors can contribute to the METHODS development and progression of the disease, including lifestyle factors (e.g., an unhealthy Delphi Methodology diet or decreased physical activity) related to obesity [2]. Developing T2D is associated with a A scientific committee of six experts defined five significant increase in the risk of cardiovascular different T2D patient profiles in ascending order (CV) and microvascular disease, but reducing of complexity (T2D patient not controlled on: levels of glycated hemoglobin (HbA1c) can sig- metformin; metformin ? dipeptidyl peptidase- nificantly decrease the risk of these complica- 4 inhibitor [DPP4-I]; metformin ? DPP4- tions [3]. To achieve and maintain long-term I ? sulfonylureas [SU]; metformin ? DPP4- metabolic control, it is necessary to combine a I ? basal insulin; metformin ? basal-bolus number of changes in the patient’s lifestyle with insulin [[ 1 IU/kg/day]) and an array of vari- different pharmacological drugs [4]. However, ables that could condition treatment choice in only about 50% of people with diabetes in Spain each profile (including body mass index [BMI], achieve their HbA1c target [5–8]. HbA1c levels, CV risk, heart failure, hyperten- A healthy lifestyle and reduction in body sion, albuminuria, estimated glomerular filtra- weight are the first therapeutic measures to be tion rate [eGFR], patient frailty, and non- considered in a patient newly diagnosed with alcoholic fatty liver disease [NAFLD]). T2D [9]. Primary pharmacologic therapy usually In the first phase of the Think Twice Pro- begins with metformin, unless otherwise con- gram, an online patient management survey of traindicated or not tolerated [4, 10, 11]. Once 145 questions matching the five patient profiles metformin fails to maintain a satisfactory and conditioning variables was developed. metabolic control, initiation of a combination Sixty-nine Spanish endocrinologists with expe- therapy should be considered [10]. However, rience in the treatment of T2D (C 5 years the key drivers influencing the endocrinolo- treating T2D, [ 50 T2D patient visits/month, gist’s choice of a second antihyperglycemic and members of the Spanish Society of Diabetes agent to be added to background therapy are [SED] and/or the Spanish Society of not well known, and the management of T2D Endocrinology and Nutrition [SEEN]) were patients in clinical practice may vary signifi- invited to participate; these specialists answered cantly. While recent position statements and the survey according to their clinical practice guidelines consider obesity, risk of hypo- and gathered in local meetings to discuss the glycemia, and CV disease to be critical factors results. for the selection of a second agent [4, 9, 10], In a second phase, the scientific committee T2D heterogeneity, complex patient profiles developed 55 statements presenting treatment Diabetes Ther (2019) 10:1893–1907 1895 Table 1 Statements that achieved consensus in the Delphi study Statement Statement Median Median Participants in Consensus Reasoning behind number* (P25-P75) range median range, n (%) consensus Overall statements S2 In a frail patient on treatment with 9 (8-9) 7-9 119 (88.8) C-A Tailoring treatment is SU, repaglinide or insulin therapy, important in patients for the objectives of glycemic control whom benefits of an are set >7.5%. intensive antidiabetic treatment are limited, such as elderly patients [45]. S4 If the patient presents with 7 (7-8) 7-9 107 (79.9) C-A uncontrolled HT, I preferentially Efficacy of SGLT2-I in choose a treatment regimen reducing SBP [46] and CV including an SGLT2-I. outcomes such as S7 If the patient presents with HF, I 8 (7-9) 7-9 114 (85.1) C-A hospitalization for HF [22- choose a treatment regimen 24]. including an SGLT2-I. S10 Usually, when I start a patient on 2 (1-4) 1-3 91 (67.9) C-D** Basal insulin is the most basal insulin, I discontinue the convenient initial insulin glucose lowering drugs, except for regimen. The general metformin. recommendation is to add it to metformin/oral agents or to background GLP1-RA for a better postprandial control [4]. S11 When I add an SGLT2-I to the 2 (1-4) 1-3 95 (70.9) C-D The natural history of T2D treatment of a poorly controlled demands a proggessive patient on a treatment regimen that intensification of therapy in includes a DPP4-I, I discontinue the an add-on basis rather than DPP4-I. replacing background drugs. Case 1: T2D patient uncontrolled on metformin S12 If the patient presents with HbA1c 8 (6-9) 7-9 93 (69.4) C-A** <8.5% and BMI between 25 and Neutral effect on weight <30, I add an SGLT2-I to the showed by DPP4-Is and treatment over a DPP4-I. higher potency S13 If the patient presents with HbA1c 8 (7-9) 7-9 117 (87.3) C-A demonstrated by SGLT2-I ≥8.5% and BMI between 25 and in reducing HbA1c and <30, I add an SGLT2-I to the body weight [46-48]. treatment over a DPP4-I. S15 If the patient presents with HbA1c 3 (2-5) 1-3 71 (53.0) NC-D ≥8.5% and BMI ≥30, I add an SGLT2-I to treatment over a GLP1- RA. S15b If the patient presents with HbA1c 7 (5-8) 7-9 77 (76.2) C-A GLP1-RAs are financed by ≥8.5% and BMI ≥30, I add a the Spanish Health System GLP1-RA to treatment over an when BMI is >30 kg/m2. SGLT2-I. S17 If the patient is a frail, elderly 8 (8-9) 7-9 128 (95.5) C-A DPP4-Is do not increase the person, I add a DPP4-I to treatment. risk of hypoglycemia and are a convenient option in the fragile population [45]. S18 If the patient presents with eGFR 3 (2-5) 1-3 81 (60.4) NC-D between 30 and <60 mL/min/1.73 m2 and is obese, I add a DPP4-I to treatment over a GLP1-RA. S18b If the patient presents with eGFR 8 (7-8) 7-9 84 (83.2) C-A Approved indication of between 30 and <60 mL/min/1.73 dulaglutide and liraglutide m2 and is obese, I add a GLP1-RA in eGFRs up to 15 to treatment over a DPP4-I. mL/min/1.73 m2 [33,34]. 1896 Diabetes Ther (2019) 10:1893–1907 Table 1 continued S19 If the patient presents with eGFR 8 (7-9) 7-9 103 (76.9) C-A DPP4-Is have shown a <30 mL/min/1.73 m2, I substitute favorable safety profile in metformin with a DPP4-I. CKD [49,45]. Case 2: T2D patient uncontrolled on metformin + DPP4-I S20 If the patient presents with HbA1c 8 (7-9) 7-9 101 (75.4) C-A <8.5% and BMI between 25 and <30, I add an SGLT2-I to High potency demonstrated treatment. by SGLT2-I in reducing S22 If the patient presents with HbA1c 8 (7-9) 7-9 117 (87.3) C-A HbA1c and body weight ≥8.5% and BMI between 25 and [46] <30, I add an SGLT2-I to treatment. S24 If the patient presents with HbA1c 7 (6-8) 7-9 79 (59.0) NC-A <8.5% and BMI ≥30, I add an 7 (6-8) 7-9 74 (73.3) C-A SGLT2-I to treatment. Both SGLT2-Is and GLP1- S25 If the patient presents with HbA1c 8 (7-9) 7-9 106 (79.1) C-A RAs are known for their <8.5% and BMI ≥30, I substitute weight lowering effects [50- the DPP4-I with a GLP1-RA. 52], and one or the other are S26 If the patient presents with HbA1c 7 (6-8) 7-9 81 (60.4) NC-A recommended when needing ≥8.5% and BMI ≥30, I add an 7 (6-8) 7-9 69 (68.3) C-A** to minimize weight gain or SGLT2-I to treatment. to promote weight loss S27 If the patient presents with HbA1c 8 (8-9) 7-9 115 (85.8) C-A [17,53]. ≥8.5% and BMI ≥30, I substitute the DPP4-I with a GLP1-RA. S28 If the patient presents with HbA1c 8 (8-9) 7-9 116 (86.6) C-A Synergic effect of GLP1- ≥8.5% and BMI ≥30, I substitute RA+SGLT2-I on both the DPP4-I with a GLP1-RA and glucose and weight later I add on an SGLT2-I. reduction [54]. S30 If the patient presents with eGFR 8 (7-9) 7-9 106 (79.1) C-A Same as S18b. between 30 and <60 mL/min/1.73 m2 and BMI ≥30, I substitute the DPP4-I with a GLP1-RA over adding basal insulin to treatment. Case 3: T2D patient uncontrolled on metformin + SU + DPP4-I S32 If the patient presents with HbA1c 8 (6-9) 7-9 93 (69.4) C-A ** <8.5% and BMI between 25 and <30, I substitute the SU with an SGLT2-I. S35 If the patient presents with HbA1c 7 (6-8) 7-9 85 (63.4) NC-A <8.5% and BMI ≥30, I substitute 7 (6-8) 7-9 75 (74.3) C-A The replacement of an SU the SU and the DPP4-I with a by other antihyperglycemic combination of SGLT2-I+GLP1- agents is a growing RA. tendency in all clinical S37 If the patient presents with HbA1c 8 (7-9) 7-9 102 (76.1) C-A scenarios. ≥8.5% and BMI ≥30, I substitute the SU and the DPP4-I with a combination of SGLT2-I+GLP1- RA. Case 4: T2D patient with BMI ≥30, uncontrolled on metformin + DPP4-I + basal insulin S39 If the patient presents with HbA1c 7 (6-8) 7-9 93 (69.4) C-A** Same as S24-S27. <8.5%, I add an SGLT2-I to treatment. S40 If the patient presents with HbA1c 7 (6-9) 7-9 96 (71.6) C-A <8.5%, I substitute the DPP4-I with a GLP1-RA. Diabetes Ther (2019) 10:1893–1907 1897 Table 1 continued S41 If the patient presents with HbA1c 8 (7-8) 7-9 104 (77.6) C-A ≥8.5%, I substitute the DPP4-I with a GLP1-RA. S42 If the patient presents with HbA1c 8 (7-9) 7-9 103 (76.9) C-A ≥8.5%, I substitute the DPP4-I with a GLP1-RA and add on an SGLT2- I. S45 If the patient presents with eGFR 7 (6-8) 7-9 90 (67.2) C-A ** Same as S18b. between 30 and <60 mL/min/1.73 2 m , I substitute the DPP4-I with a GLP1-RA. Case 5: T2D patient with BMI ≥30, uncontrolled on metformin + DPP4-I + insulin basal-bolus (>1 IU/kg/day) S48 If the patient presents with HbA1c 8 (7-9) 7-9 104 (77.6) C-A <8.5%, I add an SGLT2-I to treatment. S49 If the patient presents with HbA1c 7 (7-9) 7-9 105 (78.4) C-A ≥8.5%, I add an SGLT2-I to Same as S24-S27. treatment. S51 If the patient presents with HbA1c 8 (7-9) 7-9 103 (76.9) C-A ≥8.5%, I add a GLP1-RA to treatment. S statement, SU sulfonylurea, HT hypertension, SGLT2-I sodium/glucose cotransporter-2 inhibitor, SBP systolic blood pressure, CV cardiovascular, HF heart failure, GLP1-RA glucagon-like peptide-1 receptor agonist, DPP4-I dipeptidyl peptidase-4 inhibitor, T2D type 2 diabetes, HbA1c glycated hemoglobin, BMI body mass index, eGFR estimated glomerular filtration rate, CKD chronic kidney disease, C consensus, NC non-consensus, A agreement, D disagreement Clear grey font: First round results for those statements that achieved consensus on the second round Black font: Consensus, either on first or second round Italics: Statements reworded for the second round *According to the order in which the statements were presented in the Delphi questionnaire **Consensus close to the limit (66.6%) options for complex T2D profiles that were online questionnaire by an extended panel of based on the main treatment patterns and endocrinologists (each original panelist could controversies identified in the local meetings invite up to two additional endocrinologists (Tables 1, 2). Statements were divided into six with proven experience in the treatment of T2D blocks, with a block representing each of the [same criteria as those described above]). All five previously defined T2D profiles and an endocrinologists were assigned an alphanu- additional block of common practices identified meric access code to the questionnaire website across all patient profiles. To be established as to maintain their anonymity. recommendations, consensus needed to be Statements which did not achieve consensus reached on the statements in accordance with in the first round were submitted to a subse- the Delphi methodology [12]. Each statement quent round. was rated on a Likert-like scale from 1 (‘‘com- pletely disagree’’) to 9 (‘‘completely agree’’). Statistical Analysis Responses were grouped by tertiles in which 1–3 indicated disagreement, 4–6 indicated indeter- A descriptive analysis of all items using the minate and 7–9 indicated agreement. Consen- mean ± standard deviation, median (p25–p75), sus on a statement was reached when the and minimum and maximum values was per- responses of two thirds or more participants formed. The Kolmogorov–Smirnov test was (C 66.6%) were located in the same tertile as the used to check for ‘‘goodness’’ of fit of the data to median value of all the reported responses for a normal distribution. that statement. Statements were rated on an 1898 Diabetes Ther (2019) 10:1893–1907 Table 2 Statements that did not achieve consensus in the Delphi study Statement Statement Median Median Participants in median Consensus number* (P25-P75) range range, n (%) Overall statements S1 BMI is the main factor influencing my choice of antidiabetic 7 (5-7) 7-9 71 (53.0) NC-A treatment 7 (6-7) 7-9 56 (55.4) NC-A S3 If the patient presents with albuminuria, I choose SGLT2-I 7 (5-8) 7-9 71 (53.0) NC-A treatment over GLP1-RA treatment. 7 (6-8) 7-9 61 (60.4) NC-A S5 In my clinical practice, I do not differentiate between high 4 (2-7) 4-6 24 (17.9) NC-I CV risk and established CVD for making therapeutic decisions. S5b For making therapeutic decisions, I do not differentiate 6 (3-7) 4-6 25 (24.8) NC-I between high CV risk and secondary prevention. S6 If the patient presents with established CVD, I choose 6 (5-7) 4-6 60 (44.8) NC-I SGLT2-I treatment over GLP1-RA treatment. 6 (5-7) 4-6 45 (44.6) NC-I S8 If the patient presents with NAFLD, I choose pioglitazone 5 (3-6) 4-6 60 (44.8) NC-I treatment over GLP1-RA treatment. 5 (2.5-6) 4-6 40 (39.6) NC-I S9 In poorly controlled patients on a combination of three non- 4 (2-6) 4-6 51 (38.1) NC-I insulin drugs, I prefer non-insulin quadruple therapy to 5 (3-6) 4-6 36 (35.6) NC-I adding insulin. Case 1: T2D patient uncontrolled on metformin S14 If the patient presents with HbA1c <8.5% and BMI ≥30, I 5 (3-7) 4-6 47 (35.1) NC-I add an SGLT2-I to treatment over a GLP1-RA. 6 (4-7) 4-6 36 (35.6) NC-I S16 If the patient presents with HbA1c ≥8.5% and BMI ≥30, I 7 (5-8) 7-9 85 (63.4) NC**-A use the combination GLP1-RA+SGLT2-I as add-on 7 (5-8) 7-9 57 (56.4) NC-A treatment. Case 2: T2D patient uncontrolled on metformin + DPP4-I S21 If the patient presents with HbA1c <8.5% and BMI between 4 (2-7) 4-6 41 (30.6) NC-I 25 and <30, I substitute the DPP4-I with an SGLT2-I. 6 (3.5-7) 4-6 32 (31.7) NC-I S23 If the patient presents with HbA1c ≥8.5% and BMI between 2 (1-5) 1-3 83 (61.9) NC-D 25 and <30, I substitute the DPP4-I with an SGLT2-I. 5 (2-7) 4-6 24 (23.8) NC-I S29 If the patient is frail and elder, I add basal insulin to 6 (5-8) 4-6 134 (35.8) NC-I treatment. 7 (5-8) 7-9 54 (53.5) NC-A 2 S31 If the patient presents with eGFR <30 mL/min/1.73 m , 6 (3-8) 4-6 26 (19.4) NC-I HbA1c <8.5% and BMI ≥30, I withdraw metformin and 6 (4-8) 4-6 31 (30.7) NC-I substitute the DPP4-I with a GLP1-RA over adding basal insulin to treatment. Case 3: T2D patient uncontrolled on metformin + SU + DPP4-I S33 If the patient presents with HbA1c ≥8.5% and BMI between 7 (5-8) 7-9 73 (54.5) NC-A 25 and <30, I substitute the SU with an SGLT2-I and I add 7 (5-8) 7-9 53 (52.5) NC-A basal insulin to treatment. S34 If the patient presents with HbA1c ≥8.5% and BMI between 5 (3-7) 4-6 40 (29.9) NC-I 25 and <30, I add basal insulin to treatment. 5 (3-7) 4-6 39 (38.6) NC-I S36 If the patient presents with HbA1c ≥8.5% and BMI ≥30, I 6 (4-7) 4-6 59 (44.0) NC-I substitute the SU with an SGLT2-I and add basal insulin to 6 (5-8) 4-6 42 (41.6) NC-I treatment. S38 If the patient presents with HbA1c ≥8.5% and BMI ≥30, I 4 (2-6) 4-6 45 (33.6) NC-I add basal insulin to treatment. 4 (3-6) 4-6 49 (48.5) NC-I Case 4: T2D patient with BMI ≥30, uncontrolled on metformin + DPP4-I + basal insulin S43 If the patient presents with HbA1c ≥8.5%, I intensify the 4 (2-6) 4-6 51 (38.0) NC-I insulin therapy with prandial insulin. 3 (2-6) 1-3 52 (51.5) NC-D S44 If the patient is a frail, elderly person, I intensify insulin 4 (2-5) 4-6 46 (34.3) NC-I therapy with prandial insulin. 3 (2-6) 1-3 51 (50.5) NC-D Diabetes Ther (2019) 10:1893–1907 1899 Table 2 continued S46 If the patient presents with eGFR between 30 and <60 4 (2-6) 4-6 66 (49.2) NC-I mL/min/1.73 m2, I intensify insulin therapy with prandial 4 (3-6) 4-6 41 (40.6) NC-I insulin. 2 S47 If the patient presents with eGFR <30 mL/min/1.73 m , I 7 (5-8) 7-9 68 (50.7) NC-A discontinue metformin and intensify insulin therapy with 6 (4-8) 4-6 41 (40.6) NC-I prandial insulin. Case 5: T2D patient with BMI ≥30, uncontrolled on metformin + DPP4-I + insulin basal-bolus (>1 IU/kg/day) S50 If the patient presents with HbA1c ≥8.5%, I substitute the 6 (4-8) 4-6 41 (30.6) NC-I insulin bolus with GLP1-RA. 6 (5-8) 4-6 34 (33.7) NC-I S52 If the patient presents with eGFR between 30 and <60 6 (4-8) 4-6 43 (32.1) NC-I 2 mL/min/1.73 m , I substitute the insulin bolus with GLP1- 6 (5-8) 4-6 38 (37.6) NC-I RA. S53 If the patient presents with eGFR <60 mL/min/1.73 m2, I 7 (3-8) 7-9 68 (50.7) NC-A discontinue metformin and add a GLP1-RA on to treatment. 7 (5-8) 7-9 57 (56.4) NC-A 2 S54 If the patient presents with eGFR <30 mL/min/1.73 m , I 7 (5-8) 7-9 76 (56.7) NC-A discontinue metformin and intensify insulin treatment. 6 (4-7) 4-6 40 (39.6) NC-I S55 If the patient presents with eGFR <30 mL/min/1.73 m2, I 5 (2-7) 4-6 49 (36.6) NC-I discontinue metformin and add a DPP4-I to treatment. 5 (3-7) 4-6 37 (35.2) NC-I S statement, BMI body mass index, SGLT2-I sodium/glucose cotransporter-2 inhibitor, GLP1-RA glucagon-like peptide-1 receptor agonist, CV cardiovascular, CVD cardiovascular disease, NAFLD non-alcoholic fatty liver disease, T2D type 2 diabetes, HBA1c glycated hemoglobin, DPP4-I dipeptidyl peptidase-4 inhibitor, eGFR estimated glomerular filtration rate, SU sulfonylurea, C consensus, NC non-consensus, A agreement, D disagreement, I indeterminate Clear grey font: First round results Black font: Second round results Italics: Statements reworded for the second round *According to the order in which the statements were presented in the Delphi questionnaire **Close to the limit of achieving consensus (66.6%) The internal consistency of the question- k = 0.61–0.80 indicating good agreement, and naire was measured by the Cronbach’s alpha k = 0.81–1 indicating very good agreement [16]. (Ca), which can range between 0 and 1, from Statistics were calculated for the overall survey lower to higher reliability (acceptable val- and for each of the six blocks. Statistical signif- ues: [ 0.7; high reliability: 0.7–0.9; very high icance was considered when p \ 0.05. reliability [ 0.9) [13]. In addition, inter-rater The coefficient of variation (COV) of the reliability was assessed by the intra-class corre- questionnaire was calculated for every round, lation coefficient (ri) (poor: ri \ 0.40; fair: along with the delta or relative change in the ri = 0.40–0.59; good: ri = 0.60–0.74; excellent: second round above the first (COVsecond – ri = 0.75–1.0) [14]. Correlation between the two COVfirst/COVfirst). When the absolute value of rounds of the questionnaire was measured by delta is B 10%, there is no large variability the Spearman coefficient (rs) (none or poor: between the rounds and, thus, there is no need rs = 0–0.25; weak: rs = 0.26–0.50; moderate to for another round, since no relevant changes strong: rs = 0.51–0.75; and strong to very are expected. strong: rs = 0.76–1) [15]. The Kappa index (k) was calculated to estimate the qualitative Compliance with Ethics Guidelines agreement between rounds of the question- naire, taking into account the three response This study is based on a clinical practice ques- groups (1–3, 4–6 and 7–9), with k \ 0.20 indi- tionnaire that does not involve the participa- cating no or poor qualitative agreement, tion of human subjects or patient data k = 0.21–0.40 indicating weak agreement, management and does not aim to modify the k = 0.41–0.60 indicating moderate agreement, current clinical practice of participants. As such, 1900 Diabetes Ther (2019) 10:1893–1907 this study was deemed exempt from requiring study, consensus was reached for a total of 30 ethical approval. Consent for publication of statements (agreement on 28 statements; dis- survey results was granted from all the experts agreement on two statements), while consensus participating in the program and undertaking was not reached on 25 statements (Table 2). the survey. Approved sentences and reasoning underlying each consensus are summarized in Table 1. Overall, regardless of patient profile, a RESULTS sodium-glucose transport protein 2 inhibitor (SGLT2-I) is added to the treatment of patients Of the 192 endocrinologists from all over Spain with T2D and uncontrolled hypertension (S4) who were invited to participate, 134 (70%) or heart failure (S7). answered the first round of the Delphi ques- Regarding overweight patients inadequately tionnaire. The respondents were representative controlled with metformin, treatment with an of the diverse local realities and administrations SLGT2-I is preferred over treatment with a of the autonomous communities; their mean DPP4-I (S12, S13). If the patient is already being age was 45.2 ± 10.3 years, 57.5% were women, treated with a DPP4-I, an SGLT2-I is added onto and the median number of years of experience the treatment regimen (S11, S20, S22) rather in treating patients with T2D was 17 (in- than it replacing the DPP4-I [21, 23]. Con- terquartile range [IQR] 10–27 years). Consensus versely, if the treatment regimen includes a SU, was reached for 25 of the 55 statements (agree- the SU is usually replaced by an SGLT2-I (S32). ment on 23 statements; disagreement on two Current treatment trends in uncontrolled statements). obese patients include the addition of a gluca- A second round of the Delphi questionnaire gon-like peptide-1 receptor agonist (GLP1-RA; was then performed with the 30 statements for S15) or, alternatively, an SGLT2-I (S24, S26, S39) which consensus had not been reached. State- to background therapy. If the treatment regi- ments number S5, S15, and S18 were rephrased men already includes a DPP4-I, it is replaced by for this second round for clarity. A total of 101 a GLP1-RA (S25, S27, S40, S41). When the endocrinologists (75.4% of the extended panel) HbA1c target is not reached, there was consen- answered the second round (mean age of sus to progress to triple therapy with metformin respondents was 45.0 ± 10.0 years; 55.4% were and a combination of GLP1-RA and SGLT2-I women; median years of experience in treating (S28, S42). Finally, in uncontrolled complicated patients with T2D was 19 [IQR 11–26.5 years]), obese patients already on insulin basal-bolus with consensus being reached for 5 additional therapy, either an SGLT2-I or a GLP1-RA are sentences (agreement). added to treatment (S48, S49, S51) rather than The overall questionnaire showed high substituting the insulin bolus with a GLP1-RA internal consistency (Ca) (Electronic Supple- (S50). mentary Material [ESM] Table S1). Spearman The panel sets the objectives of glycemic correlation values (rs) were moderate to very control to a target of [ 7.5% in frail patients on strong, indicating an acceptable quantitative treatment with SU, repaglinide, or insulin correlation between rounds. Kappa index therapy (S2). In elderly patients uncontrolled (k) values showed a moderate to very good on metformin, DPP4-Is remain the preferred qualitative agreement between rounds in all treatment option (S17, S19). However, in com- blocks (ESM Table S2). plex patients with an eGFR \ 60 mL/min/ The COV of the questionnaire in the first and 1.73 m2, the use of some GLP1-RAs, such as second rounds was 40 and 36%, respectively, liraglutide and dulaglutide, is displacing the use achieving a relative increase of 10%, which is of DPP4-Is (S18, S45) and even of insulin (S30). within the limits for third round consideration. A variety of statements did not reach con- However, since the absolute difference between sensus in our study, and so the situations that rounds was \ 5% (i.e., 4%), further rounds were they present remain controversial in clinical not undertaken. Thus, at the end of the Delphi practice. Diabetes Ther (2019) 10:1893–1907 1901 Fig. 1 Treatment algorithm based on the main findings sodium/glucose cotransporter-2 inhibitor, BMI body mass on statements for each patient profile. T2D Type 2 index, GLP1-RA glucagon-like peptide-1 receptor agonist, diabetes, DPP4-I dipeptidyl peptidase-4 inhibitor, SU eGFR estimated glomerular filtration rate sulfonylurea, HT hypertension, HF heart failure, SGLT2-I DISCUSSION glycocentric or adipocentric view [18] to a CV- based model where the CV safety of T2D treat- In this Delphi study, consensus was reached for ments prevails [19]. a variety of statements regarding the treatment Management of patients with T2D in pri- of patients with T2D and different levels of mary or secondary CV prevention is highly complexity, reflecting that the clinical practice heterogeneous in clinical practice (S5). Until in Spain generally follows the recommenda- recently, the main body of evidence regarding tions of the main clinical guidelines. Main the use of GLP1-RAs and SGLT2-Is came from findings on statements for each patient profile trials mainly involving secondary prevention in are depicted in the treatment algorithm shown patients ([ 60%). In these trials, treatment with in Fig. 1. both liraglutide and semaglutide showed lower Consensus was not reached on a number of rates of CV death, of nonfatal myocardial statements, mainly because of the array of infarction, or of nonfatal stroke versus placebo, treatment options available in clinical practice. along with a beneficial effect on a composite Firstly, it remains unclear whether BMI is the outcome of prespecified renal events [20, 21]. major factor influencing treatment choice (S1). Similar results were observed with empagliflozin A plausible explanation is that current clinical and canagliflozin (although it should be poin- practice involves a patient-centered approach, ted out that while GLP1-RAs mainly help by with a tailored selection of medications based reducing albuminuria, SGLT2-Is also help by on a global assessment of comorbidities, costs, preventing a decrease in the GFR), along with a and patient preferences [4, 17]. Additionally, reduction in the rates of hospitalization for the key drivers influencing treatment choice heart failure [22, 23]. Recently, the results of the have changed over time, moving from a DECLARE trial assessing the CV safety of 1902 Diabetes Ther (2019) 10:1893–1907 dapagliflozin have been published, broadening study, no final consensus was reached regarding the evidence base regarding the use of SGLT2-Is the addition of an SGLT2-I over a GLP1-RA to in primary prevention in patients. This latter the treatment of uncontrolled patients with study included [ 17,000 patients with T2D with T2D and albuminuria. Nevertheless, 60% of the either atherosclerotic CV disease (ASCVD; sec- endocrinologists supported this option (S3), ondary prevention [40%]) or multiple risk fac- probably due to the evidence supporting a tors for CV disease (CVD; primary prevention greater role for SGLT2-Is rather than for GLP1- [60%]) [24]. A significant reduction in the co- RAs in renoprotection [28]. primary composite endpoint of CV death or Regarding the use of pioglitazone in patients hospitalization for heart failure was observed, with NAFLD (S8), despite evidence of its favor- mainly due to a lower rate of hospitalization for able effect in this context and its CV benefits heart failure [24]. In addition, preliminary [29, 30], no consensus was reached, possibly due results from the REWIND trial have pointed out to concerns regarding the adverse effects of a significant reduction of major adverse CV pioglitazone in patients with comorbidities events in the subpopulation without estab- [29, 31]. lished CVD [25]. In general, when adding more than one No consensus was reached regarding the use antihyperglycemic agent to monotherapy, or of SGLT2-Is over GLP1-RAs in patients with T2D replacing two drugs from a triple therapy, and established CVD (S6). This finding may be endocrinologists follow a stepwise approach. related to logistic issues, as GLP1-RAs are not This may be the explanation for the lack of financed by the Spanish Health System in peo- consensus regarding the simultaneous addition ple with a BMI of \ 30 kg/m2. In addition, the of SGLT2-I ? GLP1-AR to metformin (S16), or current trend of splitting the definition of the replacement of SU ? DPP4-I in a patient ‘‘established CVD’’ into ASCVD and heart fail- also receiving basal insulin with this combina- ure, given the differing benefits shown by tion (S35, S37). Nevertheless, more studies on SGLT2-Is and GLP1-RAs on the prevention of the simultaneous versus sequential addition of specific CV outcomes [20–24], may have had a antidiabetic agents would be needed to broaden strong influence on treatment choice. In fact, the evidence available on this issue. the most recently updated guidelines recom- Regarding the clinical practice in more mend an SGLT2-I or a GLP1-RA for the man- complex patients uncontrolled on a combina- agement of patients with T2D and established tion of three non-insulin drugs, no consensus ASCVD, while an SGLT2-I is preferred for was reached on what to do next (S9). Interest- patients at risk of heart failure. A GLP1-RA is ingly, when given the option to add insulin to only recommended for this latter group of the therapeutic program of these patients, 24% patients when an SGLT2-I is not tolerated or of the endocrinologists preferred a non-insulin contraindicated, or the eGFR is less than ade- quadruple therapy despite the absence of quate [10, 17]. Head-to-head studies regarding specific recommendations in the main guideli- CV outcomes comparing SGLT2-Is and GLP1- nes, reflecting that insulin is the last resort for RAs could add supporting evidence on this some physicians [32]. Similar results were issue. observed for uncontrolled patients on a thera- As previously mentioned, most of the peutic program of metformin ? DPP4-I ? basal SGLT2-I CV outcome trials have reported con- insulin and HbA1c levels of C 8.5%, where the sistent benefits in renal endpoints, including addition of an SGLT2-I or the replacement of albuminuria [23, 24, 26, 27]. In fact, both the the DPP4-I by a GLP1-RA (± SGLT2-I) prevailed American Diabetes Association (ADA) and the over intensifying the insulin therapy with European. Association for the Study of Diabetes prandial insulin (S41–S43), confirming the (EASD) recommend an SGLT2-I in patients with apparent preference of Spanish endocrinologists T2D and chronic kidney disease, with or with- for non-insulin therapies. out CVD, or the use of a GLP1-RA if the SGLT2-I No consensus was reached regarding the use is contraindicated or not preferred [17]. In our or intensification of insulin in frail patients Diabetes Ther (2019) 10:1893–1907 1903 (S29, S44). Likewise, there was no agreement on in the local meetings of the Think Twice pro- intensifying therapy with prandial insulin in gram, a full list of whom are listed in ESM 3. patients non-controlled with met- formin ? basal insulin ? DPP4-I and with an Funding. Logistic support, including the eGFR between 30 and \ 60 mL/min/1.73 m2 journal’s Rapid Service fee, was provided by (S46), or in more complex patients with an ESTEVE Pharmaceuticals S.A Spain. All authors eGFR of \ 30 mL/min/1.73 m2 (S47), possibly had full access to all of the data in this study because the panel preferred the option of and take complete responsibility for the integ- replacing the DPP4-I with a GLP1-RA (S45). This rity of the data and accuracy of the data preference may be partly explained by the reg- analysis. ulatory approval of dulaglutide and liraglutide in eGFRs up to 15 mL/min/1.73 m2 [33, 34], Medical Writing and Editorial Assis- together with the greater effectiveness demon- tance. Eva Pineda at ESTEVE Medical Affairs strated by GLP1-RAs in reducing weight and provided scientific support. Almudena Pardos hypoglycemic episodes in comparison to pran- and Anna Nualart at Ogilvy Health provided dial rapid-acting insulins [35–38]. Nevertheless, medical writing services. This assistance was these studies may not reflect the complexity of funded by ESTEVE Pharmaceuticals S.A Spain. patients in clinical practice, as they included patients initially treated only with metformin Authorship. All named authors meet the and/or SU [37–43] who therefore had a poten- International Committee of Medical Journal tially better beta-cell function, which is associ- Editors (ICMJE) criteria for authorship for this ated with better glycemic responses to GLP1-RA article, take responsibility for the integrity of therapy [44]. the work as a whole, and have given their approval for this version to be published. CONCLUSIONS Disclosures. Carlos Morillas has participated in advisory boards organized by Sanofi, Novo In this Delphi study, consensus was reached on Nordisk, Lilly, Boehringer, Janssen, Esteve, and a number of statements regarding the manage- Astra-Zeneca; and has acted as a speaker for MSD, ment of patients with T2D and different levels Astra-Zeneca, Novartis, Sanofi, Novo Nordisk, of complexity. These statements reflect the Boehringer, Esteve, Janssen, and Lilly. Javier most frequent behavior in current clinical Escalada has participated in advisory boards practice of Spanish endocrinologists in the organized by Lilly, MSD, Novo Nordisk, Sanofi, treatment of T2D, and may be used as guidance and Esteve, and has presented data in events for other physicians, especially in those sce- organized by Almirall, Astra-Zeneca, Boehringer, narios where the evidence available is scarce. Janssen, Lilly, MSD, Novo Nordisk, and Sanofi. Achieving a more homogeneous management Rafael Palomares has taken part in advisory of complex T2D profiles may require further boards for Boehringer, Novartis, Sanofi-Aventis, evidence on the strengths and weaknesses of Novo Nordisk, Abbot, and AstraZeneca Pharma- therapies currently available and a better ceuticals LP; and has acted as a speaker for Astra- understanding of the key drivers influencing Zeneca, Abbott, Almirall, Bayer, Boehringer, treatment choice. Bristol-Myers, Esteve, Faes Farma, Ferrer, GSK, Janssen, Lacer, Lifescan, Lilly, Menarini, Merck, MSD, Novartis, Novo Nordisk, Pfizer, Roche, ACKNOWLEDGEMENTS Sanofi-Aventis, Servier and Takeda. Diego Bellido has taken part in advisory boards for Sanofi, The authors would like to thank the panel of Boehringer Lilly, and Novo Nordisk; and has endocrinologists for their participation in the acted as a speaker for MSD, Lilly, Novo Nordisk, Delphi study. The authors would also like to AstraZeneca, Esteve, and Sanofi. Fernando extend their special thanks to the participants Gómez-Peralta has taken part in advisory panels 1904 Diabetes Ther (2019) 10:1893–1907 for Sanofi and Novo Nordisk; has received 2. Bellou V, Belbasis L, Tzoulaki I, Evangelou E. Risk research support from Sanofi, Novo Nordisk, factors for type 2 diabetes mellitus: An exposure- wide umbrella review of meta-analyses. PLoS One. Boehringer Ingelheim Pharmaceuticals, and 2018;13(3):e0194127. https://doi.org/10.1371/ Lilly; and has acted as a speaker for Sanofi, Novo journal.pone.0194127. Nordisk, Boehringer Ingelheim Pharmaceuticals, AstraZeneca Pharmaceuticals LP, and Lilly. 3. Stratton IM, Adler AI, Neil HA, et al. Association of glycaemia with macrovascular and microvascular Antonio Pérez has served as a consultant for or complications of type 2 diabetes (UKPDS 35): received lecture fees or travel reimbursement prospective observational study. BMJ. from Sanofi Aventis, Almirall, Novo Nordisk, Eli 2000;321(7258):405–12. Lilly, MSD, Boehringer Ingelheim, Esteve, Bris- 4. Association of American Diabetes. 9. Pharmacologic tol-Myers Squibb, Novartis, Amgen, Menarini, approaches to glycemic treatment: standards of and Astra Zeneca. medical care in diabetes-2019. Diabetes Care. 2019;42[Suppl 1]:S90–102. https://doi.org/10.2337/ Compliance with Ethics Guidelines. This dc19-s009. study is based on a clinical practice questionnaire 5. Vinagre I, Mata-Cases M, Hermosilla E, et al. Con- that does not involve the participation of human trol of glycemia and cardiovascular risk factors in subjects nor patient data management and does patients with type 2 diabetes in primary care in not aim to modify the current clinical practice of Catalonia (Spain). Diabetes Care. 2012;35(4):774–9. https://doi.org/10.2337/dc11-1679. participants. Consequently, as per ethical approval procedures in Spain and the Council of 6. Perez A, Franch J, Cases A, et al. Relationship Europe’s classification of this article as a scientific between the degree of glycemic control and dia- audit, the questionnaires compiled in this study betes characteristics and hyperglycemia treatment in type 2 diabetes. DIABES Study. Med Clin (Barc). did not require ethical approval. Consent for 2012;138(12):505–11. https://doi.org/10.1016/j. publication of survey results was granted from all medcli.2011.06.026. the experts participating in the program and undertaking the survey. 7. Perez A, Mediavilla JJ, Minambres I, Gonzalez-Se- gura D. Glycemic control in patients with type 2 diabetes mellitus in Spain. Rev Clin Esp. Data Availability. The datasets during and/ 2014;214(8):429–36. https://doi.org/10.1016/j.rce. or analyzed during the current study are avail- 2014.05.028. able from the corresponding author on reason- able request. 8. Minambres I, Mediavilla JJ, Sarroca J, Perez A. Meeting individualized glycemic targets in primary care patients with type 2 diabetes in Spain. BMC Open Access. This article is distributed Endocr Disord. 2016;16:10. https://doi.org/10. under the terms of the Creative Commons 1186/s12902-016-0090-1. Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/ 9. Garber AJ, Abrahamson MJ, Barzilay JI, et al. Con- sensus Statement by the American Association of by-nc/4.0/), which permits any noncommer- Clinical Endocrinologists and American College of cial use, distribution, and reproduction in any Endocrinology on the comprehensive type 2 dia- medium, provided you give appropriate credit betes management algorithm—2019 executive to the original author(s) and the source, provide summary. Endocr Pract. 2019;25(1):69–100. https:// doi.org/10.4158/CS-2018-0535. a link to the Creative Commons license, and indicate if changes were made. 10. Reyes-Garcia R, Moreno-Pérez O, Tejera-Pérez C, et al. Documento de Abordaje Integral de la Dia- betes Tipo 2. http://www.seen.es/docs/apartados/ 791/abordaje%20integral%20dm2019.pdf. (2019). Accessed 28 Feb 2019. REFERENCES 11. Gunton JE, Cheung NW, Davis TM, Zoungas S, 1. International Diabetes Federation. IDF diabetes Colagiuri S, Australian Diabetes Society. A new atlas. http://www.diabetesatlas.org (2017). Accessed blood glucose management algorithm for type 2 5 Oct 2018. diabetes: a position statement of the Australian Diabetes Society. Med J Aust. 2014;201(11):650–3. Diabetes Ther (2019) 10:1893–1907 1905 12. New Jersey Institute of Technology. The Delphi 24. Wiviott SD, Raz I, Bonaca MP, et al. Dapagliflozin method. Techniques and applications. Newark: and cardiovascular outcomes in type 2 diabetes. New Jersey Institute of Technology; 2002. N Engl J Med. 2018;380:347–57. https://doi.org/10. 1056/nejmoa1812389. 13. Schmitt N. Uses and abuses of coefficient alpha. Psychol Assess. 1996;8(4):350–3. 25. Gerstein HC, Colhoun HM, Dagenais GR, et al. Dulaglutide and cardiovascular outcomes in type 2 14. Cicchetti DV. Guidelines, criteria, and rules of diabetes (REWIND): a double-blind, randomised thumb for evaluating normed and standardized placebo-controlled trial. Lancet. assessment instrument in psychology. Psychol 2019;394(10193):121–30. https://doi.org/10.1016/ Assess. 1994;6(4):284–90. https://doi.org/10.1037/ s0140-6736(19)31149-3. 1040-3590.6.4.284. 26. Wanner C, Inzucchi SE, Lachin JM, et al. Empagli- 15. Dawson B, Trapp RG. Basic and clinical biostatistics flozin and progression of kidney disease in type 2 (LANGE Basic Science). 4th ed. New York: McGraw- diabetes. N Engl J Med. 2016;375(4):323–34. Hill; 2004. https://doi.org/10.1056/NEJMoa1515920. 16. Landis JR, Koch GG. The measurement of observer 27. Jardine MJ, Mahaffey KW, Neal B, et al. The cana- agreement for categorical data. Biometrics. gliflozin and renal endpoints in diabetes with 1977;33(1):159–74. established nephropathy clinical evaluation (CRE- DENCE) study rationale, design, and baseline 17. Davies MJ, D’Alessio DA, Fradkin J, et al. Manage- characteristics. Am J Nephrol. 2017;46(6):462–72. ment of hyperglycemia in type 2 diabetes, 2018. https://doi.org/10.1159/000484633. A Consensus Report by the American Diabetes Association (ADA) and the European Association 28. Muskiet MH, Tonneijck L, van Bommel EJ, Smits for the Study of Diabetes (EASD). Diabetes Care. MM, van Raalte DH. Renoprotection in LEADER 2018;41(12):2669–701. https://doi.org/10.2337/ and EMPA-REG OUTCOME. Lancet Diabetes Endo- dci18-0033. crinol. 2016;4(10):812–4. https://doi.org/10.1016/ s2213-8587(16)30214-5. 18. Gorgojo Martinez JJ. Glucocentricity or adipocen- tricity: a critical view of consensus and clinical 29. Kernan WN, Viscoli CM, Furie KL, et al. Pioglita- guidelines for the treatment of type 2 diabetes zone after ischemic stroke or transient ischemic mellitus. Endocrinol Nutr. 2011;58(10):541–9. attack. N Engl J Med. 2016;374(14):1321–31. https://doi.org/10.1016/j.endonu.2011.09.007. https://doi.org/10.1056/NEJMoa1506930. 19. Giugliano D, Maiorino MI, Bellastella G, Esposito K. 30. Bril F, Kalavalapalli S, Clark VC, et al. Response to Diabetes is a cardiovascular disease, isn’t it? Dia- pioglitazone in patients with nonalcoholic steato- betes Res Clin Pract. 2018;135:229–31. https://doi. hepatitis with vs without type 2 diabetes. Clin org/10.1016/j.diabres.2017.11.006. Gastroenterol Hepatol. 2018;16(4):558–66. https:// doi.org/10.1016/j.cgh.2017.12.001. 20. Marso SP, Daniels GH, Brown-Frandsen K, et al. Liraglutide and cardiovascular outcomes in type 2 31. Lewis JD, Habel LA, Quesenberry CP, et al. Piogli- diabetes. N Engl J Med. 2016;375(4):311–22. tazone use and risk of bladder cancer and other https://doi.org/10.1056/NEJMoa1603827. common cancers in persons with diabetes. JAMA. 2015;314(3):265–77. https://doi.org/10.1001/jama. 21. Marso SP, Bain SC, Consoli A, et al. Semaglutide 2015.7996. and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2016;375(19):1834–44. 32. Peyrot M, Rubin RR, Lauritzen T, et al. Resistance to https://doi.org/10.1056/nejmoa1607141. insulin therapy among patients and providers: results of the cross-national Diabetes Attitudes, 22. Zinman B, Wanner C, Lachin JM, et al. Empagli- Wishes, and Needs (DAWN) study. Diabetes Care. flozin, cardiovascular outcomes, and mortality in 2005;28(11):2673–9. type 2 diabetes. N Engl J Med. 2015;373(22):2117–28. https://doi.org/10.1056/ 33. European Medicines Agency. Dulaglutide. SPC. NEJMoa1504720. https://www.ema.europa.eu/documents/product- information/trulicity-epar-product-information_ 23. Neal B, Perkovic V, Matthews DR. Canagliflozin and en.pdf (2018). Accessed 3 July 2019. cardiovascular and renal events in type 2 diabetes. N Engl J Med. 2017;377(21):2099. https://doi.org/ 34. European Medicines Agency. Liraglutide. SPC. 10.1056/NEJMc1712572. https://www.ema.europa.eu/documents/product- information/victoza-epar-product-information_en. pdf (2018). Accessed 3 July 2019. 1906 Diabetes Ther (2019) 10:1893–1907 35. Abd El Aziz MS, Kahle M, Meier JJ, Nauck MA. A added to oral agents in patients with poorly con- meta-analysis comparing clinical effects of short- or trolled type 2 diabetes. Diabetes Obes Metab. long-acting GLP-1 receptor agonists versus insulin 2015;17(2):170–8. https://doi.org/10.1111/dom. treatment from head-to-head studies in type 2 dia- 12406. betic patients. Diabetes Obes Metab. 2017;19(2):216–27. https://doi.org/10.1111/dom. 44. Jones AG, McDonald TJ, Shields BM, et al. Markers 12804. of beta-cell failure predict poor glycemic response to GLP-1 receptor agonist therapy in type 2 dia- 36. Singh S, Wright EE Jr, Kwan AY, et al. Glucagon-like betes. Diabetes Care. 2016;39(2):250–7. https://doi. peptide-1 receptor agonists compared with basal org/10.2337/dc15-0258. insulins for the treatment of type 2 diabetes melli- tus: a systematic review and meta-analysis. Diabetes 45. Gomez-Huelgas R, Gomez Peralta F, Rodriguez Obes Metab. 2017;19(2):228–38. https://doi.org/10. Manas L, et al. Treatment of type 2 diabetes mellitus 1111/dom.12805. in elderly patients. Rev Clin Esp. 2018;218(2):74–88. https://doi.org/10.1016/j.rce. 37. Maiorino MI, Chiodini P, Bellastella G, Capuano A, 2017.12.003. Esposito K, Giugliano D. Insulin and glucagon-like peptide 1 receptor agonist combination therapy in 46. Storgaard H, Gluud LL, Bennett C, et al. Benefits type 2 diabetes: a systematic review and meta- and harms of sodium-glucose co-transporter 2 analysis of randomized controlled trials. Diabetes inhibitors in patients with type 2 diabetes: a sys- Care. 2017;40(4):614–24. https://doi.org/10.2337/ tematic review and meta-analysis. PLoS One. dc16-1957. 2016;11(11):e0166125. https://doi.org/10.1371/ journal.pone.0166125. 38. Castellana M, Cignarelli A, Brescia F, Laviola L, Giorgino F. GLP-1 receptor agonist added to insulin 47. Rosenstock J, Bailey CJ, Mathieu C, Chen H, Garcia- versus basal-plus or basal-bolus insulin therapy in Sanchez R, Saraiva GL. Composite endpoint analy- type 2 diabetes: a systematic review and meta- sis of dapagliflozin versus saxagliptin as add-on analysis. Diabetes Metab Res Rev. 2019;35(1):e3082. therapy in patients with type 2 diabetes inade- https://doi.org/10.1002/dmrr.3082. quately controlled with metformin. Endocr Pract. 2016;23(1):38A–9A. 39. Bunck MC, Diamant M, Corner A, et al. One-year treatment with exenatide improves beta-cell func- 48. Deacon CF. Dipeptidyl peptidase-4 inhibitors in the tion, compared with insulin glargine, in met- treatment of type 2 diabetes: a comparative review. formin-treated type 2 diabetic patients: a Diabetes Obes Metab. 2011;13(1):7–18. https://doi. randomized, controlled trial. Diabetes Care. org/10.1111/j.1463-1326.2010.01306.x. 2009;32(5):762–8. https://doi.org/10.2337/dc08- 1797. 49. Coppolino G, Leporini C, Rivoli L, et al. Exploring the effects of DPP-4 inhibitors on the kidney from 40. Russell-Jones D, Vaag A, Schmitz O, et al. Liraglu- the bench to clinical trials. Pharmacol Res. tide vs insulin glargine and placebo in combination 2018;129:274–94. https://doi.org/10.1016/j.phrs. with metformin and sulfonylurea therapy in type 2 2017.12.001. diabetes mellitus (LEAD-5 met ? SU): a randomised controlled trial. Diabetologia. 50. Manigault KR, Thurston MM. Liraglutide: a gluca- 2009;52(10):2046–55. https://doi.org/10.1007/ gon-like peptide-1 agonist for chronic weight s00125-009-1472-y. management. Consult Pharm. 2016;31(12):685–97. https://doi.org/10.4140/TCP.n.2016.685. 41. Diamant M, Van Gaal L, Guerci B, et al. Exenatide once weekly versus insulin glargine for type 2 dia- 51. Goring S, Hawkins N, Wygant G, et al. Dapagli- betes (DURATION-3): 3-year results of an open-label flozin compared with other oral anti-diabetes randomised trial. Lancet Diabetes Endocrinol. treatments when added to metformin monother- 2014;2(6):464–73. https://doi.org/10.1016/S2213- apy: a systematic review and network meta-analy- 8587(14)70029-4. sis. Diabetes Obes Metab. 2014;16(5):433–42. https://doi.org/10.1111/dom.12239. 42. Gurkan E, Tarkun I, Sahin T, Cetinarslan B, Canturk Z. Evaluation of exenatide versus insulin glargine 52. Zhang YJ, Han SL, Sun XF, et al. Efficacy and safety for the impact on endothelial functions and car- of empagliflozin for type 2 diabetes mellitus: meta- diovascular risk markers. Diabetes Res Clin Pract. analysis of randomized controlled trials. Medicine 2014;106(3):567–75. https://doi.org/10.1016/j. (Baltimore). 2018;97(43):e12843. https://doi.org/ diabres.2014.09.046. 10.1097/MD.0000000000012843. 43. D’Alessio D, Haring HU, Charbonnel B, et al. 53. Gomez-Peralta F, Escalada San Martin FJ, Menendez Comparison of insulin glargine and liraglutide Torre E, et al. Spanish Diabetes Society (SED) Diabetes Ther (2019) 10:1893–1907 1907 recommendations for the pharmacologic treatment exenatide or dapagliflozin alone in patients with of hyperglycemia in type 2 diabetes: 2018 update. type 2 diabetes inadequately controlled with met- Endocrinol Diabetes Nutr. 2018;65(10):611–24. formin monotherapy (DURATION-8): a 28 week, https://doi.org/10.1016/j.endinu.2018.08.004. multicentre, double-blind, phase 3, randomised controlled trial. Lancet Diabetes Endocrinol. 54. Frias JP, Guja C, Hardy E, et al.. Exenatide once 2016;4(12):1004–16. https://doi.org/10.1016/ weekly plus dapagliflozin once daily versus S2213-8587(16)30267-4.