sign in sign up
<<
Home , Empagliflozin

MRIGTEAIS RG N REGIMENS AND DRUGS THERAPIES: EMERGING

Diabetes Care 1

fl Julio Rosenstock,1 Jan Marquard,2 Empagli ozin as Adjunctive to Lori M. Laffel,3 Dietmar Neubacher,4 Stefan Kaspers,2 David Z. Cherney,5 Therapy inType 1 Diabetes: Bernard Zinman,6 Jay S. Skyler,7 Jyothis George,2 Nima Soleymanlou,8 and The EASE Trials Bruce A. Perkins6 https://doi.org/10.2337/dc18-1749

OBJECTIVE To evaluate the safety and efficacy of empagliflozin 10- and 25-mg doses plus a 1 unique lower dose (2.5 mg) as adjunct to intensified insulin in patients with type 1 Dallas Diabetes Research Center at Medical City, Dallas, TX diabetes (T1D). 2Boehringer Ingelheim International GmbH, Ingelheim, Germany RESEARCH DESIGN AND METHODS 3Joslin Diabetes Center, Harvard Medical School, The EASE (Empagliflozin as Adjunctive to inSulin thErapy) program (N = 1,707) Boston, MA 4 included two double-blind, placebo-controlled phase 3 trials: EASE-2 with empagli- Pharma GmbH & Co. KG, Biberach an der Riss, Germany flozin 10 mg (n = 243), 25 mg (n = 244), and placebo (n = 243), 52-week treatment; 5Division of Nephrology, Department of Medi- and EASE-3 with empagliflozin 2.5 mg (n = 241), 10 mg (n = 248), 25 mg (n = 245), cine, and Department of Physiology, Toronto and placebo (n = 241), 26-week treatment. Together they evaluated empagliflozin General Hospital, University of Toronto, Toronto, 10 mg and 25 mg, doses currently approved in treatment of , Canada 6Lunenfeld-Tanenbaum Research Institute, Mount and additionally 2.5 mg on 26-week change in (primary Sinai Hospital, and Division of Endocrinology and end point) and weight, glucose time-in-range (>70 to £180 mg/dL), insulin dose, Metabolism, University of Toronto, Toronto, , and . Canada 7Diabetes Research Institute, University of Miami RESULTS Miller School of Medicine, Miami, FL 8Boehringer Ingelheim (Canada) Ltd./Ltee,´ Bur- The observed largest mean placebo-subtracted glycated hemoglobin reductions lington, Canada were 20.28% (95% CI 20.42, 20.15) for 2.5 mg, 20.54% (20.65, 20.42) for 10 mg, Corresponding author: Julio Rosenstock, and 20.53% (20.65, 20.42) for 25 mg (all P < 0.0001). Empagliflozin 2.5/10/25 mg [email protected], or Bruce doses, respectively, reduced mean weight by 21.8/23.0/23.4 kg (all P < 0.0001); A. Perkins, [email protected]. increased glucose time-in-range by +1.0/+2.9/+3.1 h/day (P < 0.0001 for 10 and Received 16 August 2018 and accepted 20 Sep- 25 mg); lowered total daily insulin dose by 26.4/213.3/212.7% (all P < 0.0001); and tember 2018. decreased systolic blood pressure by 22.1/23.9/23.7 mmHg (all P < 0.05). Genital Clinical trial reg. nos. NCT02414958 and NCT02580591, clinicaltrials.gov. infections occurred more frequently on empagliflozin. Adjudicated diabetic keto- fl This article contains Supplementary Data online acidosis occurred more with empagli ozin 10 mg (4.3%) and 25 mg (3.3%) but was at http://care.diabetesjournals.org/lookup/suppl/ comparable between empagliflozin 2.5 mg (0.8%) and placebo (1.2%). Severe doi:10.2337/dc18-1749/-/DC1. hypoglycemia was rare and frequency was similar between empagliflozin and J.R. and J.M. contributed equally as primary placebo. coauthors. N.S. and B.A.P. contributed equally as senior CONCLUSIONS coauthors. Empagliflozin improved glycemic control and weight in T1D without increasing © 2018 by the American Diabetes Association. hypoglycemia. Ketoacidosis rate was comparable between empagliflozin 2.5 mg Readers may use this article as long as the work and placebo but increased with 10 mg and 25 mg. Ketone monitoring for early is properly cited, the use is educational and not for profit, and the work is not altered. More infor- ketoacidosis detection and intervention and lower empagliflozin doses may help mation is available at http://www.diabetesjournals to reduce this risk. .org/content/license. Diabetes Care Publish Ahead of Print, published online October 4, 2018 2 Empagliflozin as an Adjunct to Insulin in TID Diabetes Care

Type 1 diabetes (T1D), an autoimmune T1D included two international, multi- range was also selected in light of the disease characterized by insulin defi- center, phase 3, randomized, double- HbA1c superiority trial design followed in ciency, affects 30 million people world- blind, placebo-controlled, parallel-group EASE-2 and EASE-3. Key exclusion criteria wide and is associated with reduced life trials of once-daily oral empagliflozin included use of noninsulin antihypergly- expectancy owing to acute and chronic doses conducted over 52 weeks (EASE-2) cemic drugs or severe hypoglycemia or complications (1–4). The Diabetes Con- and 26 weeks (EASE-3). The treatment DKA within 3 months of inclusion. See trol and Complications Trial (DCCT) and period was preceded by a 6-week insu- Supplementary Data for the detailed list its follow-up Epidemiology of Diabetes lin intensification period and a 2-week of inclusion/exclusion criteria. Control and Complications (EDIC) study placebo run-in period and followed by have shown that improved glucose a 3-week safety follow-up. Empagliflozin Trial Procedures control by insulin intensification in T1D 10 mg and 25 mg versus placebo were Patients underwent a 6-week investigator- reduces the long-term risks of microvas- studiedinbothtrials,andanadditional guided insulin intensification period cular and macrovascular events (4). arm (empagliflozin 2.5 mg) was included that resulted in changes to HbA1c in Attaining and sustaining glycated he- in EASE-3 in order to characterize a lower addition to body weight and total daily moglobin (HbA1c) targets via insulin op- effective and safe dose (Supplemen- insulin dose from the screening visit timization strategies remains a major tary Figure 1). The design and conduct to baseline. The EASE-2/EASE-3 mean challenge owing to treatment complex- of EASE-2 and EASE-3 were identical with changes 6 SD in HbA1c, body weight, and ity, increased hypoglycemia, and poten- the exception of the following differences total daily insulin dose during this tial for weight gain. Despite advances in EASE-3: a shorter treatment duration, pretreatment period were, respec- in insulin formulations, delivery systems, the assessment of continuous glucose tively, 20.6 6 0.6%/20.5 6 0.7%, and glucose monitoring, only one-third monitoring (CGM) as a substudy, and +0.6 6 2kg/+0.56 2 kg, and +5 6 of patients are able to achieve glycemic the inclusion of a lower dose (2.5 mg). 29%/+3 6 16%. The insulin regimen targets and many become overweight In EASE-2/EASE-3, respectively, 1,338/ wastoremainstableduringasub- or obese (3,5,6). Consequently, there is 1,751 patients were screened by 131/189 sequent 2-week placebo run-in period. a need to evaluate available safe and centers across 17/24 countries; 1,015/ Trial medication was taken once daily effective treatment options to overcome 1,353 started the placebo run-in period, and adherence was evaluated at clinic suboptimal glucometabolic control in of which 730/977 were randomized with visits. If HbA1c was ,8.0% at randomiza- T1D. In this regard, the evaluation of stratification for type of insulin therapy, tion, total insulin dose was reduced by some therapies, proven to be effective estimated glomerular filtration rate 10% to lower hypoglycemia risk. During in type 2 diabetes (T2D), as adjunct to (eGFR), HbA1c, and, in EASE-3, also by the insulin intensification pretreatment insulin represents a promising strategy participation in the CGM substudy. phase and throughout the entire duration (7–9). Trial protocols and informed consent of randomized treatment, investigators Based on their insulin-independent forms were approved by institutional were unblinded to glycemic markers glucosuric mechanism, sodium–glucose review boards. Patients provided con- (e.g., fasting plasma glucose, HbA1c, cotransporter 2 inhibitors (SGLT2i) have sent prior to enrollment. Adjudication of etc.), and could freely adjust the insulin been shown in T1D clinical trials to im- cardiovascular events, severe hypoglyce- regimen according to their clinical dis- prove glucometabolic outcomes (10–13). mia, DKA, and hepatic events was per- cretion and based on local guidelines However, an increased risk of diabetic formed by masked, independent clinical to achieve the best standard of care. ketoacidosis (DKA) has raised valid clin- event committees. Trial progress and In addition, guidance to avoid substantial ical concern (14). Interestingly, previous safety were assessed by an unmasked insulin dose reduction was provided. trials in patients with T1D have tested the and independent data monitoring During the entire trial period, including same doses of SGLT2i used in T2D patients committee. Trials were sponsored by the prerandomization period, insulin despite potential differences in renal re- Boehringer Ingelheim. See Supplementary dose levels (total, basal, bolus) were sponse (15). Data for details. determined based on patient-reported Empagliflozin, a highly selective SGLT2i, information collected on a daily basis in is approved for use in adults with T2D to Trial Patients an electronic diary; data were averaged improve glycemic control and to reduce Key inclusion criteria included the fol- over a 2-week period before the time the risk of cardiovascular death (16). lowing: adult patients with eGFR $30 point of assessment. Phase 2 trials with empagliflozin have mL/min/1.73 m2,BMI$18.5 kg/m2, All patients received a point-of-care shown promise in T1D (17–24). We fasting C-peptide value ,0.7 ng/mL device capable of measuring blood present the totality of the empagliflozin (,0.23 nmol/L), T1D diagnosis $1year, glucose and b-hydroxybutyrate (BHB). phase 3 data as adjunctive to insulin in insulin needs of 0.3–1.5 units/kg on mul- Patients were educated on ketone mon- T1D including the characterization of a tiple daily injections or continuous sub- itoring when feeling unwell (e.g., illness, unique lower dose. cutaneous insulin infusion, and HbA1c symptoms suggestive of DKA irrespec- 7.5–10.0% following the lead-in insulin tive of the glucose value) and to seek intensification period. The HbA1c range medical care in case of increased BHB RESEARCH DESIGN AND METHODS of 7.5–10.0% at randomization enabled the (.1.5 mmol/L). The BHB threshold Design and Conduct inclusion of a broad population of patients of .1.5 mmol/L was chosen based on The EASE (Empagliflozin as Adjunctive to with T1D at less than optimal glycemic recommendations provided in the user inSulin thErapy) program in patients with targets despite insulin intensification. This manual of the ketone meter and in light care.diabetesjournals.org Rosenstock and Associates 3

of the fact that patients are at a higher (10 and 25 mg) and placebo using population having an eGFR ,60 mL/min/ 2 risk of developing DKA above this BHB 225 evaluable patients per arm (SD = 1.73 m . Average baseline HbA1c was level (25). During run-in and the first 0.9%). To allow for attrition, 240 patients 8.1–8.2% with insulin needs of approx- 4 weeks of treatment, fasting BHB was per arm were planned. imately 0.7 units/kg (with equal basal- tested daily to provide initial background The primary end point was analyzed bolus split). Insulin pumps were used information irrespective of symptoms, using a mixed-effects model for repeated in approximately one-third of patients. and 2–3 times/week subsequently. All measures (MMRM) with Bonferroni- Baseline characteristics were balanced patients were also provided with an adjusted comparisons between empagli- (Table 1). electronic diary for daily recording of flozin 10 or 25 mg and placebo (each dose glucose self-monitoring results, hypogly- tested at the two-sided level of a =2.5%). Primary Efficacy End Point cemic events, insulin intake, and BHB The primary efficacy analysis included Empagliflozin improved glycemic control, measurements. A masked CGM system on-treatment data only as observed as assessed by placebo-corrected HbA1c (Dexcom G4, blinded mode) was used in cases (OC) on the full analysis set (FAS), change after 26 weeks of treatment (Fig. EASE-2 in all patients (and as a substudy including all treated patients with a 1). All empagliflozin doses led to statis- fi $ in EASE-3) to assess glycemic pro le at baseline and 1 on-treatment HbA1c tically significant HbA1c reductions with baseline (over 2 weeks) and on treat- measurements. Subsequently, an effec- consistency across the trials and between ment (over 4 weeks and 2 weeks in tiveness analysis including data after the primary efficacy and effectiveness EASE-2 and EASE-3, respectively). See treatment discontinuation (OC-AD) was analyses (Fig. 1B and Supplementary Fig. fi Supplementary Data for detailed trial performed hierarchically on the modi ed 3). Maximal HbA1c effect was observed procedures. intention-to-treat set (mITT), including all from week 12 and largely sustained up treated patients with a baseline and $1 to the end of the trials (Fig. 1A and B). End Points postrandomization HbA1c measurements. Mean HbA1c reduction after 26 weeks The primary end point in both studies If efficacy and effectiveness null hypoth- of treatment was dose-dependent and was the change from baseline in HbA1c at eses were rejected, then sequentially the greatest with empagliflozin 10- and 25- week 26. Key secondary end points were primary efficacy end point for empagli- mg doses (up to 20.54%; P , 0.0001). investigator-reported symptomatic hy- flozin 2.5 mg in EASE-3 (on-treatment Empagliflozin 2.5 mg also reduced poglycemia with confirmed blood glu- data) and key secondary end points HbA1c (20.28%; P , 0.0001). The largest cose ,54 mg/dL (,3.0 mmol/L) and/or in both trials were to be tested for placebo-corrected HbA1c reduction oc- fl severe hypoglycemia requiring third- empagli ozin 10 and 25 mg versus pla- curred in patients with baseline HbA1c party assistance from weeks 5 to 26 cebo in a confirmatory way. A negative $8%, consisting of ;60% of the study as well as from weeks 1 to 26. The week binomial model was used to analyze population (2.5 mg, 20.35%; 10 mg, up 5 to 26 time window was chosen as the investigator-reported hypoglycemia. to 20.70%; 25 mg, up to 20.64%; P , first step for the key secondary hypo- AnMMRMmodelwasusedtoanalyze 0.0001), as shown in Fig. 1C. glycemia analysis in order to generate changes in weight, insulin dose, and data on a stable insulin background by bloodpressure.InEASE-2,ANCOVAwas Key Secondary Efficacy End Points excluding the initial phase of therapy used for CGM analyses. All analyses After 26 weeks of treatment, empagli- (weeks 1 to 4), when insulin dose adjust- were prespecified except for patient- flozin resulted in placebo-corrected ments are more likely to occur. In EASE- reported nocturnal hypoglycemia (Fig. reduction in body weight (up to 23.4 kg; 2, change from baseline in body weight 2B), net benefit analysis (Supplementary P , 0.0001), systolic blood pressure (up at week 26, percentage of time spent Fig. 14), and DKA subgroup analyses to 23.9 mmHg; P , 0.0001), and diastolic in target glucose range of .70 to #180 (Supplementary Table 2). See Supple- blood pressure (up to 22.3 mmHg; P = mg/dL (.3.9 to #10.0 mmol/L) and in- mentary Data for details on statistical 0.0006) with overall comparable results terquartile range (IQR) as determined by methods. for 10 and 25 mg across studies (Table 2 CGM in weeks 23 to 26, total daily insulin and Supplementary Figs. 4–6). In EASE-2, dose at week 26, and systolic/diastolic RESULTS empagliflozin doses 10 and 25 mg signif- blood pressure at week 26 were also Patient Characteristics icantly increased CGM-derived glucose evaluated as key secondary end points. Overall, 1,338 patients were screened time in range (up to 3.1 h/day; P , These parameters were also evaluated and 730 were assigned treatment in 0.0001) and decreased glycemic variabil- in EASE-3; CGM-based assessments were, EASE-2, while 1,751 were screened ity assessed by IQR (up to219 mg/dL; P , however, done in a substudy. Safety eval- and 977 were assigned treatment in 0.0001) (Table 2 and Supplementary Fig. uations consisted of adverse event (AE) EASE-3 (Supplementary Fig. 2). More 7). In EASE-3, empagliflozin 2.5 mg fol- reporting, laboratory tests, and vital signs. than 90% of patients completed week lowed the same beneficial trend as Definitions of hypoglycemia, DKA catego- 26 and were included in the full analysis the 10- and 25-mg doses with respect ries, and the AEs of interest are outlined (Supplementary Fig. 2). The study pop- to improvements in weight (21.8 kg; in Supplementary Data. ulation was half female with average P , 0.0001), systolic blood pressure baseline age of low to mid-40s, largely (22.1 mmHg; P = 0.027), glucose Statistical Analyses white, and recruited primarily in Europe time in range (+1 h/day; P =0.1063), A two-sided t test (type I error a = 2.5%) and North America. Patients had nor- and IQR (27.9 mg/dL; P = 0.1096). To- provides 90% power to detect an HbA1c mal blood pressure and kidney func- tal insulin dose was also significantly change of 20.3% between empagliflozin tion at baseline with ,4% of the overall decreasedonempagliflozin versus Empagli 4 fl zna nAjntt nui nTID in Insulin to Adjunct an as ozin Table 1—Baseline characteristics EASE-2 EASE-3 Empagliflozin 10 mg Empagliflozin 25 mg Placebo Empagliflozin 2.5 mg Empagliflozin 10 mg Empagliflozin 25 mg Placebo Characteristics (N = 243) (N = 241) (N = 239) (N = 237) (N = 244) (N = 242) (N = 238) Female 125 (51.4) 130 (53.9) 130 (54.4) 119 (50.2) 130 (53.3) 119 (49.2) 124 (52.1) Age, years 45.7 6 12.5 45.3 6 13.9 44.5 6 13.5 43.4 6 14.2 42.4 6 13.3 44.2 6 13.5 42.2 6 13.2 Race White 230 (94.7) 227 (94.2) 225 (94.1) 233 (98.3) 232 (95.1) 228 (94.2) 223 (93.7) Black/African American 6 (2.5) 4 (1.7) 8 (3.3) 4 (1.7) 10 (4.1) 4 (1.7) 5 (2.1) Asian 6 (2.5) 10 (4.1) 8 (3.3) 3 (1.3) 2 (0.8) 5 (2.1) 2 (0.8) Other 1 (0.4) 0 0 0 1 (0.4) 5 (2.1) 8 (3.4) Region Europe 131 (53.9) 130 (53.9) 133 (55.6) 156 (65.8) 148 (60.7) 150 (62.0) 148 (62.2) North America 95 (39.1) 94 (39.0) 91 (38.1) 60 (25.3) 61 (25.0) 60 (24.8) 63 (26.5) Pacific 12 (4.9) 10 (4.1) 10 (4.2) 7 (3.0) 12 (4.9) 11 (4.5) 5 (2.1) Latin America 0 0 0 12 (5.1) 10 (4.1) 14 (5.8) 19 (8.0) Africa 0 0 0 2 (0.8) 13 (5.3) 7 (2.9) 3 (1.3) Asia 5 (2.1) 7 (2.9) 5 (2.1) 0 0 0 0 Diabetes duration, years 22.8 6 12.6 22.5 6 13.0 22.4 6 12.4 20.8 6 11.9 20.5 6 11.9 21.2 6 11.4 21.7 6 13.0

HbA1c,% 8.10 6 0.60 8.06 6 0.53 8.13 6 0.57 8.14 6 0.61 8.19 6 0.64 8.19 6 0.65 8.19 6 0.58 ,8.0% 105 (43.2) 109 (45.2) 108 (45.2) 101 (42.6) 106 (43.4) 98 (40.5) 98 (41.2) $8.0% 138 (56.8) 132 (54.8) 131 (54.8) 136 (57.4) 138 (56.6) 144 (59.5) 140 (58.8) Weight, kg 86.2 6 18.2 85.6 6 18.3 83.4 6 16.7 81.6 6 14.6 83.7 6 17.0 83.3 6 18.9 80.7 6 16.9 BMI, kg/m2* 29.5 6 5.5 29.5 6 6.0 28.5 6 5.3 28.0 6 4.4 28.7 6 5.1 28.4 6 5.6 27.8 6 5.1 eGFR, mL/min/1.73 m2 95.0 6 18.1 94.1 6 18.9 95.9 6 18.4 96.5 6 19.9 97.3 6 19.9 95.7 6 19.7 97.8 6 19.3 $60 236 (97.1) 232 (96.3) 232 (97.1) 227 (95.8) 234 (95.9) 228 (94.2) 230 (96.6) ,60 7 (2.9) 9 (3.7) 7 (2.9) 10 (4.2) 10 (4.1) 14 (5.8) 8 (3.4) Blood pressure, mmHg Systolic 124.0 6 15.5 125.0 6 14.9 124.7 6 15.7 123.5 6 14.6 125.1 6 15.1 124.6 6 15.0 120.6 6 14.8 Diastolic 75.8 6 9.3 77.2 6 9.2 75.6 6 9.9 75.2 6 8.9 76.9 6 8.7 75.4 6 9.2 74.7 6 9.1 Total daily insulin dose, units/kg† 0.70 6 0.24 0.74 6 0.26 0.70 6 0.23 0.70 6 0.24 0.71 6 0.24 0.71 6 0.24 0.70 6 0.24 Basal insulin dose‡ 0.36 6 0.16 0.38 6 0.17 0.37 6 0.14 0.36 6 0.14 0.37 6 0.14 0.37 6 0.15 0.36 6 0.15 Bolus insulin dose¶ 0.34 6 0.15 0.36 6 0.17 0.35 6 0.17 0.36 6 0.17 0.35 6 0.15 0.35 6 0.15 0.35 6 0.16 Type of insulin MDI 144 (59.3) 143 (59.3) 142 (59.4) 156 (65.8) 162 (66.4) 161 (66.5) 157 (66.0) CSII 99 (40.7) 98 (40.7) 97 (40.6) 81 (34.2) 82 (33.6) 81 (33.5) 81 (34.0) Data are n (%) or means 6 SD. The eGFR at baseline was estimated according to the Epidemiology Collaboration equation. CSII, continuous subcutaneous insulin infusion; MDI, multiple daily injections. *EASE-3: n = 236 for empagliflozin 2.5 mg. †EASE-2: n = 230 for empagliflozin 10 mg, n = 226 for empagliflozin 25 mg, n = 226 for placebo; EASE-3: n = 224 for empagliflozin 2.5 mg, n = 221 for empagliflozin 10 mg, n = 226 for empagliflozin 25 mg, n = 223 for placebo. ‡EASE-2: n = 184 for empagliflozin 10 mg, n = 186 for empagliflozin 25 mg, n = 172 for placebo; EASE-3: Care Diabetes n = 192 for empagliflozin 2.5 mg, n = 190 for empagliflozin 10 mg, n = 189 for empagliflozin 25 mg, n = 187 for placebo. ¶EASE-2: n = 186 for empagliflozin 10 mg, n = 187 for empagliflozin 25 mg, n = 170 for placebo; EASE-3: n = 189 for empagliflozin 2.5 mg, n = 190 for empagliflozin 10 mg, n = 187 for empagliflozin 25 mg, n = 187 for placebo. care.diabetesjournals.org Rosenstock and Associates 5

Figure 1—HbA1c efficacy analysis: Data are from patients treated with $1 dose of study drug who had a baseline and $1 on-treatment HbA1c measurement. A: HbA1c from screening to week 52 (EASE-2) and week 26 (EASE-3). Values at screening and week 0 are descriptive data. Values at weeks 4–52 are based on mixed-model repeated measures analysis. B: Placebo-corrected change from baseline in HbA1c with empagliflozin at week 26 (primary end point in EASE-2 and EASE-3) and week 52 (EASE-2). Data are based on mixed-model repeated measures analysis. ***P , 0.0001 for difference vs. placebo. C: Placebo-corrected change from baseline in HbA1c with empagliflozin at week 26 in subgroups by baseline HbA1c (,8% and $8%). Data are based on mixed-model repeated measures analysis. **P , 0.001 for difference vs. placebo; ***P , 0.0001 for difference vs. placebo. placebo after 26 weeks of treatment: up and comparable to the level reported versus placebo after 26 weeks of treat- to 213.3% and 212.7% for empagliflozin following the end of the prerandomiza- ment (Supplementary Figs. 11 and 12). 10- and 25-mg doses, respectively, tion insulin intensification period. For and 26.4% for the 2.5-mg dose (Table patients assigned to empagliflozin, the Hypoglycemia 2 and Supplementary Fig. 8). The need placebo-corrected insulin dose reduc- Over treatment weeks 5 to 26, empagli- to reduce insulin dose when initiating tion was equivalent between basal/ flozin 2.5, 10, and 25 mg did not increase empagliflozin occurred shortly after the bolus components after 26 weeks of the rate of investigator-reported symp- start of therapy, and the total daily treatment: 20.02/20.03 units/kg for tomatic hypoglycemia (,54 mg/dL) or insulin dose was largely stabilized by 2.5 mg, up to 20.05/20.05 units/kg severe hypoglycemia (Fig. 2A). However, week 4 of treatment as assessed by for 10 mg, and up to 20.05/20.04 these investigator-reported events (clas- patient-reported insulin dose levels units/kg for 25 mg (Supplementary Figs. sified by investigators as AEs based on (Supplementary Fig. 8). Importantly, the 9 and 10). Empagliflozin also reduced their clinical review and judgment) rep- total daily insulin dose in patients as- fasting plasma glucose (up to 235.2 resented only a subset of the patient- signed to placebo was relatively sta- mg/dL; P , 0.0001) and waist circum- reported symptomatic events (,54 mg/dL) ble over the entire treatment period ference (up to 22.9 cm; P , 0.0001) captured by electronic diary (a total of 6 Empagliflozin as an Adjunct to Insulin in TID Diabetes Care

Table 2—Secondary outcomes Empagliflozin Empagliflozin Empagliflozin 2.5 mg 10 mg 25 mg P value for differences vs. placebo Weight, kg EASE-2 (26 weeks) d 22.7 23.3 ,0.0001 for both doses EASE-2 (52 weeks)* d 23.2 23.6 ,0.0001 for both doses* EASE-3 (26 weeks) 21.8* 23.0 23.4 ,0.0001 for all CGM-derived time in glucose range of .70 to #180 mg/dL, % (h/day) EASE-2 (26 weeks) d +11.9 +12.9 ,0.0001 for both doses (+2.9 h/day) (+3.1 h/day) EASE-2 (52 weeks)* d +12.2 +12.5 ,0.0001 for both doses* (+2.9 h/day) (+3.0 h/day) EASE-3 (26 weeks)† +4.3 +10.7 +7.4 ,0.0001 for 10 mg; ,0.01 for 25 mg (+1.0 h/day) (+2.6 h/day) (+1.8 h/day) CGM-derived IQR, mg/dL EASE-2 (26 weeks) d 216.9 219.0 ,0.0001 for both doses EASE-2 (52 weeks)* d 219.8 219.4 ,0.0001 for both doses* EASE-3 (26 weeks)† 27.9 214.6 210.7 ,0.01 for 10 mg; ,0.05 for 25 mg Total daily insulin dose, % EASE-2 (26 weeks) d 213.3 212.7 ,0.0001 for both doses EASE-2 (52 weeks)* d 212.0 212.9 ,0.0001 for both doses* EASE-3 (26 weeks) 26.4 29.5 212.6 ,0.0001 for all SBP/DBP, mmHg EASE-2 (26 weeks) d 22.1/21.3 23.7/22.3 SBP: ,0.05 for 10 mg*; ,0.001 for 25 mg DBP: ,0.05 for 10 mg*; ,0.001 for 25 mg EASE-2 (52 weeks)* d 23.4/21.7 24.7/21.5 SBP: ,0.01 for 10 mg*; ,0.0001 for 25 mg* DBP: ,0.05 for both doses* EASE-3 (26 weeks) 22.1/20.3 23.9/21.7 23.7/21.4 SBP: ,0.05 for 2.5 mg*; ,0.0001 for 10 mg and 25 mg DBP: ,0.01 for 10 mg*; ,0.05 for 25 mg* Data are adjusted mean differences vs. placebo in changes from baseline based on mixed-model repeated measures, except for EASE-2 CGM data at week 26, which were performed using ANCOVA. Analyses were performed in randomized patients treated with $1 dose of study drug who had a baseline and $1 on-treatment HbA1c measurement. DBP, diastolic blood pressure; SBP, systolic blood pressure. *Nominal. †Substudy.

12,790 investigator-reported events out the 2.5-mg dose, up to 37% relative to one fatal case mainly related to delayed of a total of 23,147 patient-reported placebo (Fig. 2B). DKA diagnosis and treatment (refer to events, representing 55%). For this rea- Supplementary Data for details). son, we present both investigator- and General Safety and DKA Patients with DKA generally had at patient-reported events in Fig. 2A and B, Genital infections and generally volume least one precipitating factor, such as a respectively. depletion occurred with higher fre- concomitant illness/infection or reduced During the initial phase of insulin quency with empagliflozin than placebo insulin intake (Supplementary Table 1). adjustment (weeks 1–4), the rate of in- (Table 3). Urinary tract infections, hepatic Based on baseline subgroup analyses, vestigator-reported severe and symp- events, acute renal impairment, and female sex and insulin pump use were tomatic hypoglycemic AEs (,54 mg/dL) bone fractures occurred with similar identified as important DKA risk factors was similar (Supplementary Fig. 13). frequency on empagliflozin versus pla- in pooled analyses across EASE-2 and Based on pooled safety analyses, the cebo. One minor toe amputation was EASE-3 for placebo and empagliflozin rate of adjudicated severe hypoglycemia reported on empagliflozin 2.5 mg in a 10-mg and 25-mg dose groups (Supple- was also similar between empagliflozin patient with a history of amputations and mentary Table 2). Specifically, of the and placebo (Table 3). peripheral arterial disease. DKA data 72 patients with adjudicated certain Based on the totality of all hypogly- suggested dose-dependent risk. For con- or potential DKA, 48 patients were in- cemia events reported by patients, firmed adjudicated DKA (case defini- sulin pump users, while 24 were mul- empagliflozin 10 and 25 mg significantly tion “certain”),therateinpatientson tiple daily injection users; 53 patients reduced the rate of patient-reported empagliflozin 2.5 mg was low and similar were female, while 19 were male. Of the symptomatic hypoglycemia (,54 mg/dL) to placebo (0.8%, 1.2%; respectively), while 38 patients who had both risk factors as recorded in electronic diaries up to the rate was higher in the empagliflozin (female sex and insulin pump use) and treatment week 52 (Fig. 2B). Empagliflozin 10- and 25-mg groups compared with aconfirmed adjudicated DKA event, 2.5mgalsoshowedasimilarbeneficial placebo (4.3%, 3.3%, and 1.2%, respec- 2 were in the placebo group (represent- 26-week trend in EASE-3. Nocturnal symp- tively). There were few severe DKA cases ing 1.8% of female patients on pump in tomatic hypoglycemia (,54 mg/dL) was overall with a trend toward more severe this group), 21 were in the empagliflozin alsoreducedwithempagliflozin, including cases on empagliflozin 25 mg, including 10-mg group (representing 20.4% of care.diabetesjournals.org Rosenstock and Associates 7

Figure 2—Hypoglycemia. A: Investigator-reported symptomatic hypoglycemic adverse events with blood glucose ,54 mg/dL and/or severe hypoglycemic events in weeks 5–26 (key secondary hypoglycemia end point); analyses were performed using a negative binomial model in randomized patients treated with $1 dose of study drug who had a baseline and $1 on-treatment HbA1c measurement. During the initial phase of insulin adjustments (weeks 1–4), event rates between empagliflozin and placebo were similar (Supplementary Fig. 13). B: Patient-reported symptomatic hypoglycemic events with blood glucose ,54 mg/dL (totality of all reported events by patients). EASE-3 data from baseline to week 26 are presented for empagliflozin 2.5 mg, and pooled EASE-2 and EASE-3 data from baseline to week 52 are presented for empagliflozin 10 mg and empagliflozin 25 mg. Nocturnal episodes were those with onset between 0000 h and 0559 h. Analyses were performed using a negative binomial model in randomized patients treated with $1 dose of study drug. female patients on pump in this dose hypoglycemia (Supplementary Fig. 14). loss (up to 23.4 kg), increased glucose group), and 15 were in the empagliflozin The observed clinical benefit was consis- time in range (up to +3 h/day), and 25-mg group (representing 15.2% of tent across a range of HbA1c thresholds reductions in insulin need (up to 213%) female patients on pump in this dose (Supplementary Fig. 14). and blood pressure (up to 23.9 mmHg group). for systolic) without increased severe CONCLUSIONS hypoglycemia risk provide compel- Overall Net Benefit The empagliflozin T1D program was a ling evidence of clinical benefitfor An exploratory post hoc net clinical ben- comprehensive evaluation of the bene- empagliflozininT1Dthat,aswithother fi e t analysis found that a greater pro- fit-risk profile of this SGLT2i as an ad- SGLT2i agents (10–13), needs to be fl portion of patients on empagli ozin junctive therapy to insulin. After a weighed against the increased DKA (+23–38%) relative to placebo-treated 26-week placebo-controlled randomized risk (14). patients achieved the end point that in- treatment phase, the .0.5% HbA1c re- DKA incidence, based on the T1D Ex- cluded HbA1c reduction of at least 20.3% duction with empagliflozin 10 and 25 mg change clinic registry, is about 5% per without weight gain, without occurrences is a clinically meaningful effect over in- year in adults (26). The increased DKA risk of adjudicated DKA, and without severe tensified insulin. Furthermore, weight observed in the EASE program for the 8 Empagliflozin as an Adjunct to Insulin in TID Diabetes Care

Table 3—Adverse events EASE-2 and EASE-3 pooled EASE-3 Empagliflozin 10 mg Empagliflozin 25 mg Placebo Empagliflozin 2.5 mg Placebo Event (N = 491) (N = 489) (N = 484) (N = 241) (N = 241) Any adverse event 441 (89.8) 428 (87.5) 433 (89.5) 194 (80.5) 203 (84.2) Drug-related adverse event 221 (45.0) 226 (46.2) 158 (32.6) 70 (29.0) 56 (23.2) Adverse event leading to discontinuation 29 (5.9) 18 (3.7) 14 (2.9) 8 (3.3) 2 (0.8) Serious adverse event 64 (13.0) 42 (8.6) 44 (9.1) 13 (5.4) 16 (6.6) Death 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) Adverse events of interest Event consistent with genital infection 63 (12.8) 70 (14.3) 21 (4.3) 13 (5.4) 6 (2.5) Event consistent with 47 (9.6) 41 (8.4) 41 (8.5) 13 (5.4) 11 (4.6) Event consistent with volume depletion 12 (2.4) 16 (3.3) 8 (1.7) 1 (0.4) 3 (1.2) Lower limb amputation 0 0 0 1 (0.4) 0 Bone fracture 14 (2.9) 5 (1.0) 8 (1.7) 5 (2.1) 2 (0.8) Acute renal impairment 1 (0.2) 4 (0.8) 3 (0.6) 0 0 Hepatic event 8 (1.6) 8 (1.6) 7 (1.4) 1 (0.4) 1 (0.4) Adjudicated ketoacidosis and ketosis Patients with certain ketoacidosis 21 (4.3) 16 (3.3) 6 (1.2) 2 (0.8) 3 (1.2) Patients with .1 event 0 1 0 0 0 Number of events 21 18 6 2 3 Rate per 100 patient-years 5.94 5.05 1.77 1.65 2.52 Severity of event Severe events 2 6 1 0 1 Moderate events 13 8 4 0 1 Mild events 6 4 1 2 1 Outcome of event Sequelae 0 0 1 0 0 Fatal 0 1 0 0 0 Patients with potential ketoacidosis 15 (3.1) 13 (2.7) 6 (1.2) 3 (1.2) 1 (0.4) Number of events 16 14 6 3 1 Number of mild events 16 14 6 3 1 Patients with ketosis 155 (31.6) 178 (36.4) 76 (15.7) 41 (17.0) 32 (13.3) Patients with BHB $3.8 mmol/L* 21 (13.5) 17 (9.6) 4 (5.3) 7 (17.1) 2 (6.3) Patients with cases adjudicated as unclassifiable 0 0 0 0 0 Adjudicated severe hypoglycemia Patients with any event 20 (4.1) 13 (2.7) 15 (3.1) 3 (1.2) 6 (2.5) Number of events 33 14 21 9 6 Rate per 100 patient-years 9.54 4.02 6.35 7.66 5.22 Patients with fatal events 0 0 0 0 0 Patients with nocturnal events** 10 2 6 0 2 Data are n or n (%). Data are for patients who received at least one dose of a study drug and include events that occurred during treatment or within 7 days after the last receipt of a study drug. *Percentage of patients with BHB $3.8 mmol/L is calculated based on the number of patients with ketosis. **Onset between 0000 h and 0559 h.

higher doses is also reported similarly reductions in body weight, glucose var- dose-ranging T1D study of dapagliflozin, for sotagliflozin and dapagliflozin T1D iability, blood pressure, and the notably only the T2D-approved doses have been programs (10,12). Although this risk lower DKA risk. In addition to glucome- evaluated thus far in phase 3 trials (11,31). appeared to be similar between the tabolic improvements, the HbA1c effect Our results suggest that the use of lower empagliflozin 10- and 25-mg doses, of empagliflozin 2.5 mg comes without SGLT2i doses in T1D could achieve an cases with more severe clinical character- an increased risk of severe hypoglycemia, optimal balance between safety and istics, albeit very few in number, trended a risk observed with adjunct-to-insulin efficacy. toward the 25-mg group. Interestingly, approaches using other drug classes The challenges facing patients with the observed comparable DKA risk be- (28–30). T1D are perhaps best illustrated by tween empagliflozin 2.5 mg and placebo Based on pharmacokinetic-pharmaco- data from the T1D Exchange clinic reg- suggests that lower SGLT2i doses conceiv- dynamic modeling, increased urinary istry, which highlight the unmet need ably may help to minimize this risk in T1D. glucose excretion is observed in T1D for adjunctive therapy to insulin to im- The approximate 0.3% HbA1c reduc- versus T2D (15), suggesting differences in prove and facilitate T1D management. tion with empagliflozin 2.5 mg, although renal physiology and/or glucose handling The overall mean HbA1c was 8.4% and small in magnitude, is clinically relevant (proposed model outlined in Supplementary upwards of one-third of adult patients (27), especially when taking into consid- Fig. 15). While glucosuria and glycemic were overweight or obese, clinical char- eration the totality of effects, including benefits were observed in a 2-week acteristics that are similar to the EASE care.diabetesjournals.org Rosenstock and Associates 9

populations at baseline (5). The 18-year empagliflozin 2.5-mg dose in only one of Ingelheim, Sanofi, and Novo Nordisk. D.Z.C. has follow-up data in the EDIC study further the two phase 3 studies (up to 26 weeks) received honoraria from Boehringer Ingelheim, fi illustrate the need to address observed are regarded as limitations in this clin- EliLilly,Merck,AstraZeneca,Sano,Merck, Mitsubishi Tanabe, AbbVie, Janssen, Bayer, suboptimal glucose control and in- ical program. An important aspect and and Prometic and has received operational 2 creased BMI (.28 kg/m ) and obesity, a strength of the EASE-2 and EASE-3 trials funding for clinical trials from Boehringer clear need in T1D patients that could was that insulin intensification during Ingelheim, Eli Lilly, Merck, Janssen, Sanofi, be partially met with empagliflozin’s the pretreatment optimization period and AstraZeneca. B.Z. has received consul- fi ting fees and honoraria from AstraZeneca, bene ts (4,32). Other than the increased and over the entire randomized treat- Boehringer Ingelheim, Eli Lilly, Janssen, Merck, DKA risk with 10- and 25-mg doses, ment phase was based on local guide- Novo Nordisk, and Sanofi.J.S.S.hasactedasan empagliflozin’s general safety profile in lines and investigator judgment as advisor to Adocia, Applied Therapeutics, As- T1D patients was comparable to its safety opposed to an enforced protocol- traZeneca, Boehringer Ingelheim, DalCor, profile as evaluated in .14,000 patients driven titration algorithm that is hard Dance Biopharm, Diavacs, Duologics, Elcelyx, Eli Lilly, Esperion, Geneuro, Ideal Life, Immu- with T2D, with demonstrated risk re- to replicate in clinical practice. The nomolecular Therapeutics, Intarcia, Intrexon/ duction of cardiovascular death (238%) 6-week insulin intensification phase, ActoBio, Kamada, Merck, Orgenesis, Sanofi, and hospitalization for whichwasbasedonreal-worldinvestigator- Servier,Tolerion,vTv,Valeritas,Viacyte,and (235%) in those with previous cardio- driven patient care, was effective and Zafgen; he has research funding from the National Institutes of Health, JDRF, and the vascular disease (16). Empagliflozin’s resulted in an approximate HbA1c reduc- fi Diabetes Research Institute Foundation; and cardiovascular and renal bene ts con- tion of 0.5% across the two studies prior he chairs the Strategic Advisory Board of the tinue to be evaluated in dedicated trials to the initiation of randomized therapy. EU INNODIA consortium and has served as a for heart failure (EMPagliflozin outcomE The totality of the EASE data, with member of the board of directors of Dexcom, in Patients with chrOnic heaRt failure adequate DKA risk mitigation and use of Intarcia, and Moerae Matrix. B.A.P. has re- [EMPEROR]) and kidney disease (EMPA- a lower (2.5 mg) dose than the doses ceived speaker honoraria from Medtronic, Johnson & Johnson, Dexcom, Insulet, Novo KIDNEY), which are enrolling patients approved for use in patients with T2D, Nordisk, AstraZeneca, Abbott, and Sanofi;has with T2D and T1D. appears to show a positive benefit-risk received research grant support from Boeh- DKA risk correlated with concomitant profile for empagliflozin in T1D. In this ringer Ingelheim, Medtronic, Novo Nordisk, illness or excessive insulin dose reduc- context, empagliflozin warrants further and the Bank of Montreal; and has served as a consultant for Boehringer Ingelheim, Novo tions (e.g., pump failure). This risk ap- consideration at a lower dose as an adjunc- Nordisk, Insulet, Sanofi, Abbott, and Neuro- pears to be higher in female patients tive therapy to insulin as it is associated Metrix. J.M., D.N., S.K., J.G., and N.S. are and with insulin pump use. If an SGLTi with clinically relevant glucometabolic employees of Boehringer Ingelheim. is to be considered in T1D, it should improvements without an apparent in- Author Contributions. B.A.P. and J.R. were the not be administered with a low carbo- creasedriskofDKAandseverehypo- global coordinating investigators of EASE-2 and EASE-3 clinical trials, respectively. All authors hydrate diet and is not to be considered glycemia in adults with T1D. took part in the analysis and interpretation of for people with history of excess alcohol data as well as the drafting and critical revision of intake or in case of a recent DKA episode. the manuscript. All authors had access to the data included in this publication and had the ultimate Adherence to optimized sick-day proto- Acknowledgments. The authors sincerely responsibility for the decision to publish this cols irrespective of glucose levels, with thank the EASE phase 3 patients who generously work. J.R. and B.A.P. are the guarantors of this volunteered their invaluable time toward this emphasis on BHB measurements and work and, as such, had full access to all the data in research, all of the EASE-2 and EASE-3 investi- temporary drug discontinuation in case the study and take responsibility for the integrity gators and site research professionals and staff of the data and the accuracy of the data analysis. of an infection or acute illness, should who took part in the conduct of these trials, be implemented (33). The risk of DKA Dr. Ona Kinduryte and Elke Schuler¨ for their must be considered in the event of supportduring the reviewof this manuscript, and References nonspecific symptoms (malaise, nau- Dr. Jens Eilbracht and Ros Swallow along with 1. International Diabetes Federation. IDF Dia- ’ sea, vomiting, anorexia, abdominal pain, the sponsor s clinical trial teams for their dili- betes Atlas, 8th edition. International Diabetes gent oversight of the operational conduct of Foundation, 2017 and excessive thirst). Patients should be these trials. 2. Livingstone SJ, Levin D, Looker HC, et al.; able to promptly assess ketones/BHB to Duality of Interest. The EASE-2 and EASE-3 Scottish Diabetes Research Network epidemiol- enable early DKA detection/intervention clinical trials were supported by Boehringer ogy group; Scottish Renal Registry. Estimated in case of symptoms, regardless of glucose Ingelheim and . J.R. has life expectancy in a Scottish cohort with type 1 fi – levels. If SGLT2i are to be used in clinical served on scienti c advisory boards and received diabetes, 2008-2010. JAMA 2015;313:37 44 honoraria or consulting fees from Eli Lilly, Novo 3. McKnight JA, Wild SH, Lamb MJ, et al. Gly- practice for T1D, additional educational Nordisk, Sanofi, Janssen, Boehringer Ingelheim, caemic control of in clinical prac- guidance with self-monitoring of ketones/ and Intarcia and has received grants/research tice early in the 21st century: an international BHB will be necessary. supportfromMerck,Pfizer,Sanofi,NovoNordisk, comparison. Diabet Med 2015;32:1036–1050 The EASE program overcame common Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, 4. Nathan DM, Bayless M, Cleary P, et al.; DCCT/ limitations in the evaluation of SGLTi, AstraZeneca, Janssen, Genentech, Boehringer EDIC Research Group. Diabetes Control and Ingelheim, Intarcia, and Lexicon. L.M.L. has Complications Trial/Epidemiology of Diabetes In- such as short trial duration, lack of CGM been a consultant for Johnson & Johnson, Eli terventions and Complications study at 30 years: data, and evaluation of a T1D-specific Lilly, Sanofi, Novo Nordisk, MannKind, Merck, advances and contributions. Diabetes 2013;62: dose, but we acknowledge that the DKA Bristol-Myers Squibb, AstraZeneca, Roche, 3976–3986 mitigation strategies may not compare Dexcom, Unomedical-ConvaTec, Insulet, and 5. Miller KM, Foster NC, Beck RW, et al.; T1D Boehringer Ingelhein and has received grant Exchange Clinic Network. Current state of type 1 with those routinely used in current support from the National Institutes of Health, diabetes treatment in the U.S.: updated data clinical practice. The lack of racial dis- JDRF, the American Diabetes Association, Helmsley from the T1D Exchange clinic registry. Diabetes tribution and the assessment of the Charitable Trust, Dexcom, Insulet, Boehringer Care 2015;38:971–978 10 Empagliflozin as an Adjunct to Insulin in TID Diabetes Care

6. WeinstockRS,Schutz-Fuhrmann¨ I,ConnorCG, threshold in patients with type 1 diabetes melli- triage: beta-hydroxybutyrate versus the urine et al.; T1D Exchange Clinic Network; DPV Initia- tus. J Clin Pharmacol 2018;58:640–649 dipstick. Diabetes Care 2011;34:852–854 tive. Type 1 diabetes in older adults: Comparing 16. ZinmanB, WannerC,LachinJM, et al.; EMPA- 26. Weinstock RS, Xing D, Maahs DM, et al.; T1D treatments and chronic complications in the REG OUTCOME Investigators. Empagliflozin, car- Exchange Clinic Network. Severe hypoglycemia United States T1D Exchange and the German/ diovascular outcomes, and mortality in type 2 and in adults with type 1 Austrian DPV registries. Diabetes Res Clin Pract diabetes. N Engl J Med 2015;373:2117–2128 diabetes: results from the T1D Exchange clinic 2016;122:28–37 17. Cherney DZ, Perkins BA, Soleymanlou N, registry. J Clin Endocrinol Metab 2013;98:3411– 7. Petrie JR. SGLT2 inhibitors in type 1 diabetes: et al. The effect of empagliflozin on arterial 3419 knocked down, but up again? Lancet Diabetes stiffness and heart rate variability in subjects 27. U.S. Department of Health and Human Ser- Endocrinol 2017;5:841–843 with uncomplicated type 1 diabetes mellitus. vices, Food and Drug Administration, Center 8. Akturk HK, Rewers A, Garg SK. SGLT inhibition: Cardiovasc Diabetol 2014;13:28 for Drug Evaluation and Research. Guidance for a possible adjunctive treatment for type 1 di- 18. Cherney DZ, Perkins BA, Soleymanlou N, industryddiabetes mellitus: developing drugs abetes. Curr Opin Endocrinol Diabetes Obes et al. Renal hemodynamic effect of sodium-glucose and therapeutic biologics for treatment and 2018;25:246–250 cotransporter 2 inhibition in patients with type 1 prevention [Internet], February 2008. Available 9. Yamada T, Shojima N, Noma H, Yamauchi T, diabetes mellitus. Circulation 2014;129:587–597 from https://www.fda.gov/downloads/Drugs/ Kadowaki T. Sodium-glucose co-transporter-2 in- 19. Cherney DZ, Perkins BA, Soleymanlou N, GuidanceComplianceRegulatoryInformation/ hibitors as add-on therapy to insulin for type 1 et al. Sodium glucose cotransport-2 inhibition Guidances/UCM071624.pdf. Accessed 15 August diabetes mellitus: systematic review and meta- and intrarenal RAS activity in people with type 1 2018 analysis of randomized controlled trials. Dia- diabetes. Kidney Int 2014;86:1057–1058 28. Ahren´ B, Hirsch IB, Pieber TR, et al.; ADJUNCT betes Obes Metab 2018;20:1755–1761 20. Famulla S, Pieber TR, Eilbracht J, et al. Glu- TWO Investigators. Efficacy and safety of liraglu- 10. Dandona P, Mathieu C, Phillip M, et al.; cose exposure and variability with empagliflozin tide added to capped insulin treatment in subjects DEPICT-1 Investigators. Efficacy and safety of as adjunct to insulin in patients with type 1 di- with type 1 diabetes: the ADJUNCT TWO random- dapagliflozin in patients with inadequately con- abetes: continuous glucose monitoring data from ized trial. Diabetes Care 2016;39:1693–1701 trolled type 1 diabetes (DEPICT-1): 24 week a 4-week, randomized, placebo-controlled trial 29. Mathieu C, Zinman B, Hemmingsson JU, results from a multicentre, double-blind, phase (EASE-1). Diabetes Technol Ther 2017;19:49–60 et al.; ADJUNCT ONE Investigators. Efficacy 3, randomised controlled trial. Lancet Diabetes 21. Perkins BA, Cherney DZ, Partridge H, et al. and safety of added to insulin treat- Endocrinol 2017;5:864–876 Sodium-glucose cotransporter 2 inhibition and ment in type 1 diabetes: the ADJUNCT ONE treat- 11. Garg SK, Henry RR, Banks P, et al. Effects of glycemic control in type 1 diabetes: results of an to-target randomized trial. Diabetes Care 2016; sotagliflozin added to insulin in patients with 8-week open-label proof-of-concept trial. Dia- 39:1702–1710 type 1 diabetes. N Engl J Med 2017;377:2337– betes Care 2014;37:1480–1483 30. Younk LM, Mikeladze M, Davis SN. Pramlin- 2348 22. Perkins BA, Cherney DZ, Soleymanlou N, tide and the treatment of diabetes: a review of 12. Buse JB, Garg SK, Rosenstock J, et al. Sotagli- et al. Diurnal glycemic patterns during an the data since its introduction. Expert Opin flozin in combination with optimized insulin 8-week open-label proof-of-concept trial of Pharmacother 2011;12:1439–1451 therapy in adults with type 1 diabetes: the empagliflozin in type 1 diabetes. PLoS One 2015; 31. Henry RR, Rosenstock J, Edelman S, et al. North American inTandem1 study. Diabetes 10:e0141085 Exploring the potential of the SGLT2 inhibitor Care 2018;41:1970–1980 23. Pieber TR, Famulla S, Eilbracht J, et al. dapagliflozin in type 1 diabetes: a randomized, 13. Danne T, Cariou B, Banks P, et al. HbA1c and Empagliflozin as adjunct to insulin in patients double-blind, placebo-controlled pilot study. Di- hypoglycemia reductions at 24 and 52 weeks with type 1 diabetes: a 4-week, randomized, abetes Care 2015;38:412–419 with sotagliflozin in combination with insulin in placebo-controlled trial (EASE-1). Diabetes 32. Aiello LP; DCCT/EDIC Research Group. Dia- adults with type 1 diabetes: the European in- Obes Metab 2015;17:928–935 betic retinopathy and other ocular findings in Tandem2 study. Diabetes Care 2018;41:1981– 24. Shimada A, Hanafusa T, Yasui A, et al. the Diabetes Control and Complications Trial/ 1990 Empagliflozin as adjunct to insulin in Japanese Epidemiology of Diabetes Interventions and Com- 14. Nathan DM. Adjunctive treatments for participants with type 1 diabetes: results of plications study. Diabetes Care 2014;37:17–23 type 1 diabetes. N Engl J Med 2017;377:2390– a 4-week, double-blind, randomized, placebo- 33. Laffel LM, Wentzell K, Loughlin C, Tovar A, 2391 controlled phase 2 trial. Diabetes Obes Metab Moltz K, Brink S. Sick day management using 15. Mondick J, Riggs M, Kaspers S, Soleymanlou 2018;20:2190–2199 blood 3-hydroxybutyrate (3-OHB) compared N, Marquard J, Nock V. Population pharmaco- 25.AroraS,HendersonSO,LongT,MenchineM. with urine ketone monitoring reduces hospital kinetic-pharmacodynamic analysis to character- Diagnostic accuracy of point-of-care testing for visits in young people with T1DM: a randomized ize the effect of empagliflozin on renal glucose diabetic ketoacidosis at emergency-department clinical trial. Diabet Med 2006;23:278–284