HbA_1c and Hypoglycemia Reductions at 24 and 52 Weeks With Sotagliflozin in Combination With Insulin in Adults With Type 1 Diabetes: The European inTandem2 Study

Most adults with type 1 diabetes (T1D) have HbA_1c ≥7.0% and are overweight or obese, putting them at risk for diabetes complications (1–3). Approximately 85% report hypoglycemia, at a rate of ∼75 events/patient-year (4,5). The incidence of diabetic ketoacidosis (DKA) has reached 263 events/1,000 person-years (6). DKA incidence tends to decline during adulthood, but the incidence of severe hypoglycemia increases, especially in patients >65 years old (1,6,7).

Adjuncts to insulin could potentially address these unmet needs, but most, such as pramlintide, incretin therapies, and metformin, have shown few, if any, benefits (1,8–13). Recent 6-month studies of sodium–glucose cotransporter (SGLT) inhibitors in T1D suggest they may be safe and effective in combination with insulin (14,15). Sotagliflozin (LX4211) is a novel dual inhibitor of SGLT1 and SGLT2. Local SGLT1 inhibition in the proximal intestine reduces glucose reabsorption, thereby blunting and delaying postprandial hyperglycemia, whereas systemic SGLT2 inhibition decreases renal glucose reabsorption (16–19). In the global inTandem3 study, which involved 1,402 patients with T1D, sotagliflozin significantly reduced HbA_1c, weight, and systolic blood pressure (SBP) without increasing severe hypoglycemia, but it was associated with an increased incidence of DKA (14). Similar results were found in a 24-week study of the selective SGLT2 inhibitor dapagliflozin in T1D (15).

Because HbA_1c outcomes and acute diabetes complications may differ substantially among regions (2–4), we recruited adults with T1D from 96 European and Israeli sites to study the efficacy and safety of sotagliflozin versus placebo in combination with insulin for 52 weeks. Insulin optimization was started 6 weeks before initiating combination therapy in order to identify incremental effects of sotagliflozin that could not be achieved by merely increasing insulin doses.

This phase 3 multicenter, randomized, double-blind, placebo-controlled, parallel-group study evaluated the safety and efficacy of once-daily oral sotagliflozin 200 or 400 mg in combination with insulin in adults with T1D who had inadequate glycemic control (clinical trial reg. no NCT02421510, clinicaltrials.gov). Double-blind treatment continued for 52 weeks, and the primary end point was assessed at week 24 (Fig. 1). Randomization was stratified according to HbA_1c level at screening and use of multiple daily insulin injections (MDI) or continuous subcutaneous insulin infusion (CSII). A subgroup of patients underwent blinded continuous glucose monitoring (CGM) with a Dexcom G4 monitor (Dexcom Inc., San Diego, CA) during specified 1-week intervals throughout the first 24 weeks.

Insulin therapy optimization began 6 weeks before randomization and continued throughout the trial (Fig. 1). Study personnel adjusted basal and bolus insulin doses to maintain fasting or preprandial blood glucose between 4.4 and 7.2 mmol/L (80 and 130 mg/dL) and 1- to –2-h postprandial glucose (PPG) at <10 mmol/L (<180 mg/dL) (Supplementary Data, Insulin Dose Algorithms). An independent insulin dose monitoring committee (IDMC), comprising diabetologists and certified diabetes educators, blindly reviewed insulin titration decisions starting 6 weeks before baseline through week 24 to determine whether insulin adjustment was consistent with self-monitored blood glucose (SMBG) patterns.

The institutional review board at each study center or the local ethics committee approved the protocol, consent form, and associated documents. All patients provided written informed consent. An independent clinical end point committee, blinded to trial treatment, adjudicated episodes of severe hypoglycemia, episodes of DKA, major adverse cardiovascular events, drug-induced liver injuries, and deaths. An independent data monitoring committee reviewed safety. An independent statistician performed statistical analyses.

The study included men and nonpregnant women aged 18 years or older who had T1D treated with insulin delivered through MDI or CSII and who had HbA_1c between 7.0 and 11.0% at screening. Patients with β-hydroxybutyrate (BHB) >0.6 mmol/L at screening were excluded (see the Supplementary Data for all inclusion and exclusion criteria).

After the 6-week insulin optimization phase, patients were randomly assigned at a 1:1:1 ratio to oral sotagliflozin 200 mg, oral sotagliflozin 400 mg, or placebo, all of which were administered once daily for 52 weeks. All study medications were given as two tablets before the first meal of the day. Sotagliflozin 200 mg was given as one sotagliflozin 200-mg tablet and one placebo tablet; the 400-mg dose was given as two sotagliflozin 200-mg tablets.

Bolus insulin was reduced by 30% for the first meal after the first dose of study medication on day 1 only (20). Thereafter and throughout the 52-week study, investigators adjusted insulin doses according to SMBG results, and the IDMC evaluated the doses until week 24. In addition to premeal testing to determine bolus insulin doses, patients were asked to test blood glucose before breakfast, before and 2 h after lunch, and before dinner and bedtime (five-point profile) at least three times a week before a clinic visit. An eight-point profile (2 h after breakfast and dinner and at 2 or 3 a.m., in addition to the five-point profile) was recommended once during the week before a clinic visit. HbA_1c and laboratory fasting plasma glucose (FPG) results were masked to study staff from the start of the 6-week lead-in through week 24 (after week 12, HbA_1c values >11% were unmasked to allow appropriate intervention). After week 24, HbA_1c and laboratory FPG values were unmasked to investigators. From baseline to week 12, antihypertensive treatment was not adjusted unless required for patient safety.

All study participants received urine ketone strips, point-of-care blood BHB meters and strips, and information on urogenital hygiene, proper hydration, and detecting and treating ketosis. Study centers received recommendations for ketosis and DKA diagnosis and management (see Supplementary Data).

The primary end point was the placebo-adjusted change in HbA_1c from baseline to week 24. The first secondary end point was a composite of the proportion of patients with HbA_1c <7.0% (<53 mmol/mol), no episode of severe hypoglycemia, and no episode of DKA at week 24. Additional key secondary end points were the changes from baseline to week 24 in body weight, daily bolus insulin dose, FPG, and scores on the Diabetes Treatment Satisfaction Questionnaire status and the two-item Diabetes Distress Screening Scale. Other objectives included examining the change in SBP and diastolic blood pressure (DBP) in the entire study population, and the change in SBP in those with SBP ≥130 mmHg at baseline, from baseline to week 12 (per protocol, antihypertensive therapies could be adjusted thereafter). HbA_1c, FPG, daily insulin dose, frequency of SMBG-documented hypoglycemic events (glucose ≤3.9 and ≤3.0 mmol/L [≤70 and ≤55 mg/dL]), and kidney function were assessed at each study visit through 52 weeks, and composite end points evaluating the proportions of patients meeting HbA_1c targets without severe hypoglycemia, DKA, or weight gain were calculated at weeks 24 and 52. The CGM substudy did not meet enrollment targets (n = 70 patients per arm). Therefore, before investigators and staff were unblinded to data, the inTandem1 (conducted in the U.S. and Canada with an identical study design) and inTandem2 protocols were modified to include a prespecified pooled analysis of CGM data, the primary end point of which was the percentage of readings in the target range of 3.9–10.0 mmol/L (70–180 mg/dL).

Safety and tolerability were evaluated throughout the study. DKA diagnosis was based on evidence of metabolic acidosis and other criteria, as reviewed and adjudicated centrally by a blinded events committee (see Supplementary Data).

Efficacy analyses were based on the modified intent-to-treat population, which included all randomized patients who had taken at least one dose of the study drug. Primary efficacy end point data were analyzed using mixed-effects model for repeated measures (MMRM) statistics based on the restricted maximum likelihood method for estimation. The analysis model included fixed categorical effects of treatment, randomization strata based on use of MDI or CSII, HbA_1c ≤8.5% or >8.5% at week −2, and other covariates. MMRM or ANCOVA was used for continuous secondary and other efficacy end points, with the corresponding end point and baseline value (including first-order interactions in the MMRM) in the model. A Cochran-Mantel-Haenszel test, stratified by the randomization stratification factors, was used for binary end points. The treatment groups were compared at week 24, and descriptive statistics were provided for each clinic visit through week 52. Missing observations at week 24 were imputed as nonresponse. See the Supplementary Data for additional details.

Between May 2015 and June 2017, at 96 sites in 17 European countries and Israel, 800 patients entered the placebo run-in, and 782 were randomly assigned to treatment with placebo (n = 258), sotagliflozin 200 mg (n = 261), or sotagliflozin 400 mg (n = 263); 236, 239, and 240 patients, respectively, completed 24 weeks of treatment (Supplementary Fig. 1). During the 28-week extension period, 13 patients in each sotagliflozin group and 11 in the placebo group discontinued participation. Baseline characteristics were similar between groups (Supplementary Table 1). In the total cohort, 581 patients (74.3%) used MDI and 201 patients (25.7%) received CSII (Supplementary Table 1).

Mean HbA_1c at screening was 8.42% (68.6 mmol/mol), 8.35% (67.8 mmol/mol), and 8.38% (68.1 mmol/mol) in the placebo and sotagliflozin 200- and 400-mg groups, respectively. During 6 weeks of insulin optimization before randomization, HbA_1c decreased to baseline levels of 7.79% (61.6 mmol/mol), 7.74% (61.1 mmol/mol), and 7.71% (60.8 mmol/mol), respectively. At baseline, 44 patients (17.1%) in the placebo group, 50 patients (19.2%) in the sotagliflozin 200-mg group, and 46 patients (17.5%) in the sotagliflozin 400-mg group had HbA_1c <7.0%.

After 24 weeks, the difference from placebo for the primary end point of HbA_1c, as measured by the least squares mean (LSM) difference, was −0.37% (95% CI −0.48 to −0.25) and −0.35% (−0.47 to −0.24) with sotagliflozin 200 and 400 mg (both P < 0.001 vs. placebo). LSM differences between sotagliflozin and placebo were maintained at 52 weeks: sotagliflozin 200 mg −0.21% (−0.35 to −0.07; P = 0.003); sotagliflozin 400 mg −0.32% (−0.46 to −0.18; P < 0.001) (Fig. 2A and Supplementary Table 2). Across the entire study population, 15.1% of the placebo group, 33.3% of the sotagliflozin 200-mg group, and 33.8% of the sotagliflozin 400-mg group achieved HbA_1c <7.0% at 24 weeks; at 52 weeks the proportions were 15.5%, 27.2%, 27.8%, respectively. Among the subgroup of patients with HbA_1c ≥7.0% at baseline, HbA_1c <7.0% was achieved after 24 weeks by 13 of 214 (6.1%), 52 of 211 (24.6%; P < 0.001 vs. placebo), and 57 of 217 (26.3%; P < 0.001) patients receiving placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. After 52 weeks, 8.8%, 20.2%, and 21.6% of patients with HbA_1c ≥7.0% at baseline in the placebo, sotagliflozin 200-mg, and sotagliflozin 400-mg groups had HbA_1c <7.0%.

FPG decreased significantly in both sotagliflozin groups, with placebo-adjusted differences of −1.20 mmol/L (−21.6 mg/dL [95% CI −1.79 to −0.61 mmol/L]; P < 0.001) and −1.42 mmol/L (−25.7 mg/dL [−2.01 to −0.84 mmol/L]; P < 0.001) for sotagliflozin 200 and 400 mg, respectively, at 24 weeks. At 52 weeks, FPG was reduced by 0.27 mmol/L (4.9 mg/dL [−0.92 to 0.38 mmol/L]; P = 0.41) and 0.87 mmol/L (15.8 mg/dL [−1.52 to −0.23 mmol/L]; P = 0.008) relative to placebo (Supplementary Table 2).

The inTandem2 CGM subgroup included 48, 45, and 49 patients randomized to placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. Glycemic variability, measured on the basis of the mean amplitude of glycemic excursions and CGM SD, improved with both sotagliflozin doses (Fig. 3 and Supplementary Table 3). Relative to placebo, patients treated with sotagliflozin 200 and 400 mg spent, respectively, 8.4% (P = 0.044) and 13.4% (P < 0.001) more time with glucose values in the target range 3.9–10.0 mmol/L (70–180 mg/dL) and had lower PPG: −2.80 mmol/L (P = 0.009) and −4.20 mmol/L (P < 0.001) (Supplementary Table 3).

Insulin optimization continued in all three treatment groups. Sotagliflozin was associated with reductions in total daily insulin dose that stabilized by week 4 (Fig. 2B and Supplementary Table 2). At 24 weeks, the LSM percentage differences from placebo in daily bolus insulin were −12.94% (95% CI −20.50 to −5.38; P < 0.001) with sotagliflozin 200 mg and −16.37% (−23.90 to −8.83; P < 0.001) with sotagliflozin 400 mg. Daily basal insulin dose decreased by 5.82% (P = 0.007) and 4.67% (P = 0.03) with sotagliflozin 200 and 400 mg, respectively, relative to placebo). P values were <0.05 for placebo-adjusted percentage changes in total, bolus, and basal insulin at 52 weeks, except for daily bolus insulin in the sotagliflozin 200-mg group (Supplementary Table 2 and Supplementary Fig. 2A).

Body weight decreased significantly with sotagliflozin treatment throughout the study (Fig. 2C). Sotagliflozin 200 mg was associated with placebo-adjusted LSM weight reductions of 1.98 kg (2.48% [95% CI −2.53 to −1.44 kg]) and 2.18 kg (2.78% [−2.88 to −1.48 kg]) at 24 and 52 weeks, respectively (both P < 0.001). Mean weight loss in the sotagliflozin 400-mg group was −2.58 kg (−3.08% [−3.12 to −2.04 kg]) at 24 weeks and −2.92 kg (−3.50% [−3.62 to −2.22 kg]) at 52 weeks relative to the placebo group (both P < 0.001).

After 12 weeks, patients receiving sotagliflozin 400 mg experienced SBP reductions, with an LSM difference from placebo of −2.8 mmHg (95% CI −4.6 to −1.1 mmHg; P = 0.001). The placebo-adjusted difference in DBP was not significant with sotagliflozin 400 mg, and sotagliflozin 200 mg was not associated with significant changes in either SBP or DBP (Supplementary Table 2). Among patients with SBP ≥130 mmHg at baseline, sotagliflozin 400 mg reduced SBP by 4.1 mmHg (−7.8 to −0.5 mmHg; P = 0.027) relative to placebo at 12 weeks.

Over 52 weeks, total cholesterol increased by 0.07 mmol/L with placebo and 0.3 mmol/L in the sotagliflozin groups, and LDL cholesterol increased by 0.05 with placebo, 0.14 with sotagliflozin 200 mg, and 0.13 mmol/L with sotagliflozin 400 mg. Other lipid changes were small and not considered clinically meaningful.

In the sotagliflozin arms, an initial decrease in estimated glomerular filtration rate of >3 mL/min/1.73 m² in the sotagliflozin groups at week 4 resolved by week 24 (placebo −1.70 mL/min/1.73 m²; sotagliflozin 200 mg −1.66 mL/min/1.73 m²; sotagliflozin 400 mg −1.39 mL/min/1.73 m²) (Supplementary Fig. 3 and Supplementary Table 2).

At 24 weeks, 15.12%, 31.42%, and 32.32% of patients receiving placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively, achieved HbA_1c <7.0% without experiencing severe hypoglycemia or DKA (Supplementary Fig. 4). Treatment differences at week 24 were 16.30% (95% CI 8.79–23.82; P < 0.001) for sotagliflozin 200 mg and 17.20% (9.67–24.73; P < 0.001) for sotagliflozin 400 mg versus placebo; differences at week 52 were 11.33% (4.13–18.52; P = 0.001) and 12.27% (5.05–19.50; P < 0.001) for the sotagliflozin 200- and 400-mg doses, respectively, versus placebo (Supplementary Fig. 4). The proportions of sotagliflozin-treated patients who achieved HbA_1c <7.0% but also experienced severe hypoglycemia (<2%) or DKA (<1%) were not statistically different from those receiving placebo (≤1%) throughout the study (Supplementary Fig. 4). At 24 and 52 weeks, 44.87% and 37.26% of patients receiving sotagliflozin 400 mg achieved an HbA_1c reduction ≥0.5% without severe hypoglycemia or DKA, compared with 17.44% and 19.38% receiving placebo, respectively (Supplementary Fig. 4).

When HbA_1c and weight effects at 24 weeks were evaluated together, 25.67% of patients treated with sotagliflozin 200 mg and 29.66% of those treated with sotagliflozin 400 mg achieved HbA_1c <7.0% while not gaining weight, versus 6.98% of patients receiving placebo (Supplementary Fig. 5); at 52 weeks, the respective proportions were 19.92%, 22.81%, and 8.53% (Supplementary Table 4). More patients who achieved HbA_1c <7.0% lost >5% of body weight at 24 weeks while taking sotagliflozin 200 mg (difference from placebo, 6.12% [95% CI 2.32–9.91%]; P < 0.001) or sotagliflozin 400 mg (8.34% [4.18–12.51%]; P < 0.001) (Supplementary Table 4). Similar patterns were observed at 52 weeks, and for patients with an HbA_1c reduction ≥0.5% at 24 and 52 weeks (Supplementary Fig. 6 and Supplementary Table 4).

A composite of HbA_1c <7.0%, no weight gain, no severe hypoglycemia, and no DKA was achieved at 24 weeks by 24.14% of patients treated with sotagliflozin 200 mg and 28.14% of those treated with sotagliflozin 400 mg compared with 6.98% of those in the placebo group (Supplementary Fig. 5); the respective proportions were 18.39%, 21.67%, and 7.75% at 52 weeks (Supplementary Table 4).

Diabetes Treatment Satisfaction Questionnaire status scores increased significantly from baseline to week 24 in both sotagliflozin groups but did not change significantly in the placebo group, resulting in differences of 2.0 (95% CI 1.3–2.7; P < 0.001) in the sotagliflozin 200-mg group and 1.7 (1.0–2.4; P < 0.001) in the sotagliflozin 400-mg group (Supplementary Fig. 7 and Supplementary Table 2). At week 24, distress among sotagliflozin-treated patients, as measured by the two-item Diabetes Distress Screening Scale, decreased by 0.3 (−0.6 to 0.0; P = 0.025) and 0.4 (−0.7 to −0.2; P = 0.003) with the 200- and 400-mg doses relative to placebo. Distress values among patients in the placebo group did not change from baseline. At week 52, LSM differences from placebo were −0.2 (−0.5 to 0.1; P = 0.23) for sotagliflozin 200 mg and −0.3 (−0.6 to 0.0; P = 0.046) for sotagliflozin 400 mg.

Documented hypoglycemia at thresholds of ≤3.0 mmol/L (≤55 mg/dL) and ≤3.9 mmol/L (≤70 mg/dL) decreased in both sotagliflozin groups relative to the placebo group (Supplementary Table 2). Between weeks 23 and 24, 16.7% and 15.4% of the sotagliflozin 200- and 400-mg groups, respectively, had SMBG ≤3.0 mmol/L, compared with 28.9% of the placebo group (relative risk vs. placebo: 0.579 [95% CI 0.410–0.817]; P = 0.002 for sotagliflozin 200 mg; 0.527 [0.369–0.753]; P < 0.001 for sotagliflozin 400 mg). LSM differences in events per patient per day were −0.028 (−0.047 to −0.009; P = 0.004) and −0.031 (−0.049 to −0.012; P = 0.001) at 24 weeks, and −0.026 (−0.048 to −0.005; P = 0.017) and −0.030 (−0.051 to −0.009; P = 0.006) at 52 weeks, with the sotagliflozin 200- and 400-mg doses, respectively, compared with placebo. Sotagliflozin treatment was also associated with a lower frequency of SMBG ≤3.9 mmol/L and with lower hypoglycemic event rates (Supplementary Table 2).

Over 52 weeks, 32 of 782 patients (4.1%) experienced one or more positively adjudicated severe hypoglycemia event (13 of 258 patients [5.0%] in the placebo group, 13 of 261 patients [5.0%] in the sotagliflozin 200-mg group, and 6 of 263 patients [2.3%] in the sotagliflozin 400-mg group) (Table 1), at rates of 0.07, 0.08, and 0.06 events/patient/year, respectively. No patients discontinued participation because of severe hypoglycemia.

Over 52 weeks, 56 patients had serious or nonserious acidosis-related adverse events, and 15 of 782 patients (1.9%) had one or more positively adjudicated DKA event: none receiving placebo, 6 of 261 (2.3%) receiving sotagliflozin 200 mg (one of whom used CSII), and 9 of 263 (3.4%) receiving sotagliflozin 400 mg (five of whom used CSII) (Table 1 and Supplementary Table 5). During the first 24 weeks, two patients (0.8%) treated with sotagliflozin 200 mg and four patients (1.5%) treated with sotagliflozin 400 mg experienced one or more positively adjudicated DKA event. Over 52 weeks, four patients taking sotagliflozin 400 mg discontinued participation because of DKA (Table 1). Mean increases in BHB of ∼0.1 mmol/L were observed in the sotagliflozin groups (Supplementary Fig. 9).

During 52 weeks of double-blind treatment, adverse events were reported by 61.2%, 68.2%, and 68.8% of patients treated with placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively; acidosis-related events accounted for the differences between treatment groups (Table 1 and Supplementary Table 5). Serious adverse events occurred in more sotagliflozin-treated than placebo-treated patients (sotagliflozin 200 mg: 26 [10.0%]; sotagliflozin 400 mg: 21 [8.0%]; placebo: 17 [6.6%]). Two deaths occurred in the placebo group: one due to cardiopulmonary failure and one as the result of a lung neoplasm. Major adverse cardiovascular events occurred in one patient receiving placebo, three patients receiving sotagliflozin 200 mg, and three patients receiving sotagliflozin 400 mg. Nine patients (3.5%) receiving placebo, 10 (3.8%) receiving sotagliflozin 200 mg, and 18 (6.8%) receiving sotagliflozin 400 discontinued participation because of an adverse event (including four who discontinued as a result of DKA).

Most adverse events were mild to moderate in severity. Diarrhea (consistent with SGLT1 inhibition) occurred in 12 patients (4.6%) taking sotagliflozin 200 mg and 19 patients (7.2%) taking sotagliflozin 400 mg, compared with 9 patients (3.5%) taking placebo (Table 1). The severity of symptoms was generally mild to moderate, and only two patients (0.8%) from each sotagliflozin group and one placebo-treated patient (0.4%) discontinued participation. Twenty-four patients (9.2%) receiving sotagliflozin 200 mg and 29 (11.0%) receiving sotagliflozin 400 mg reported a genital mycotic infection (consistent with SGLT2 inhibition), compared with six patients (2.3%) receiving placebo (Table 1). Two patients each from the sotagliflozin 400-mg and placebo groups and three patients from the sotagliflozin 200-mg group discontinued participation because of a genital mycotic infection. Urinary tract infections were reported by 13 patients (5.0%), 11 patients (4.2%), and 18 patients (6.8%) receiving placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. Two patients (0.8%) from the sotagliflozin 200-mg group discontinued due to urinary tract infections. Bone fractures occurred in eight patients (3.1%) receiving placebo, six patients (2.3%) receiving sotagliflozin 200 mg, and five patients (1.9%) receiving sotagliflozin 400 mg. In the sotagliflozin 200-mg group, one patient had a toe amputated; this patient had a history of amputation of multiple toes.

Over 52 weeks of treatment, sotagliflozin 200 and 400 mg combined with an optimized insulin regimen significantly reduced HbA_1c compared with placebo and demonstrated clinically meaningful, improved outcomes beyond HbA_1c reduction, as recently recommended by Agiostratidou et al. (21). Sotagliflozin improved elements of efficacy beyond HbA_1c, such as CGM time in the target range of 3.9–10.0 mmol/L (70–180 mg/dL), glycemic variability, FPG, PPG, weight, SBP, and documented hypoglycemia events. These improvements were achieved with reductions in insulin dose. Because the changes beyond HbA_1c (notably body weight reduction and glycemic variability) were dose dependent, the sotagliflozin 400-mg dose may be appropriate for patients who would benefit from additional efficacy. Although the incidence of DKA with sotagliflozin was low and within the range reported among the general T1D population, there was an imbalance in DKA that favored placebo, as no DKA occurred in that group (1,6). At 24 and 52 weeks, clinically meaningful composite end points were achieved with sotagliflozin. At 1 year, ∼1 in 5 sotagliflozin-treated patients achieved HbA_1c <7.0% without severe hypoglycemia, DKA, or weight gain, compared with 1 in 13 patients treated with placebo. These targets were achieved in a population whose insulin doses were optimized throughout the trial, and they are consistent with findings in the inTandem3 study, in which 24% and 7% of patients treated with sotagliflozin or placebo, respectively, on a background of nonoptimized insulin achieved the same goal. Patients receiving sotagliflozin also reported statistically significant and clinically meaningful increases in satisfaction with treatment and decreases in diabetes distress.

These results are consistent with the previously published results from inTandem3, in which sotagliflozin demonstrated an ability to significantly reduce HbA_1c, weight, and blood pressure, with a reduced rate of hypoglycemia occurrence and a low incidence of DKA when combined with insulin to treat T1D (14). The inTandem1 study, a 52-week trial conducted in the U.S. and Canada with a design identical to that of the current study, showed similar results (22).

In general, higher HbA_1c at baseline is associated with more robust HbA_1c reductions when using oral agents. In Dapagliflozin Evaluation in Patients With Inadequately Controlled Type 1 Diabetes (DEPICT-1), a 24-week study of dapagliflozin in T1D in which no patients with HbA_1c <7.0% were randomized, HbA_1c decreased from 8.5% at baseline to 8.0% at week 24; the percentage of patients achieving HbA_1c <7.0% was not reported (15). By contrast, at baseline in the current study, 18% of randomized patients had HbA_1c <7.0%, and mean HbA_1c was 7.8%. A 24-week decrease to HbA_1c of 7.4% with sotagliflozin 400 mg was maintained through 52 weeks. In this population, whose insulin therapy was optimized, ∼25% of sotagliflozin-treated patients whose HbA_1c was ≥7.0% at baseline maintained a value <7.0% after 24 weeks (20% met this goal after 52 weeks), compared with ≤9% of the placebo group. Similar results were seen in the inTandem3 study, in which twice as many sotagliflozin- as placebo-treated patients achieved HbA_1c <7% with background insulin that was not optimized (14). In the general population with T1D, ≥70% of patients have HbA_1c ≥7.0% (1,2). A treatment that could help up to a quarter of these individuals achieve glucose levels that are closer to normal without experiencing hypoglycemia could be a major advance in therapy, taking into account that no other pharmacological options other than insulin regimen modifications are available to improve diabetes control in patients with T1D. Moreover, satisfactory measures for mitigating risk for DKA during SGLT inhibitor therapy are available.

CGM substudy data demonstrated PPG reductions and reductions in glycemic variability that endured throughout 24 h and showed a nearly 3-h/day increase in time in the target range of 3.9–10.0 mmol/L (70–180 mg/dL) with sotagliflozin 400 mg. The pattern of insulin dose reductions associated with these changes differed from those seen in DEPICT-1. With dapagliflozin 10 mg, total, basal, and bolus insulin doses decreased by 13%, 14%, and 18%, respectively, relative to placebo (15). With sotagliflozin 400 mg, total, basal, and bolus insulin decreased by 9%, 5%, and 16% after 24 weeks and by 8%, 7%, and 12% after 52 weeks, respectively. During the second half of this study, investigators had access to HbA_1c values and made insulin titration decisions without input from the IDMC, which may have affected outcomes. The more robust reductions in bolus insulin than in basal insulin doses in the current study compared with those in DEPICT-1 are consistent with the mechanism of SGLT1 inhibition, which delays and reduces glucose absorption in the proximal intestine and could help stabilize glucose levels by blunting PPG excursions. The modest reductions in basal insulin over 52 weeks are consistent with changes in bolus insulin, driving decreases in total daily insulin dose. These changes are hypothesized to contribute, at least in part, to lower observed rates of severe and documented hypoglycemia.

Insulin dose reduction is a risk factor for ketosis and could result in DKA (23,24). One hypothesis to be tested is whether modest versus large reductions in basal insulin dose could translate into lower risk for ketosis and, by extension, lower risk for DKA. Over the first 6 months, 1–2% of patients treated with sotagliflozin experienced a positively adjudicated DKA event. Similar rates with narrower adjudication criteria were reported for dapagliflozin (15,25). In this study, the majority of patients resumed therapy after temporary discontinuation of sotagliflozin because of a DKA event, which suggests that DKA risk is manageable. However, further work needs to be done to elucidate the risk factors for DKA. When SGLT inhibitors are administered, monitoring for ketosis is required, particularly during metabolically stressful situations. SGLT inhibitors should be temporarily discontinued before scheduled surgical procedures, and patients and clinicians should consult closely regarding other forms of behavioral and physiological stress (26).

Significant improvements in patient-reported outcomes were observed with sotagliflozin compared with placebo. These improvements may be related to lower HbA_1c and better glycemic control, because the measures of patient outcomes reflected distress at the inability to control glucose. Weight loss of 2–3 kg (3–4%) while taking sotagliflozin was also meaningful, as attempts to improve glycemic control with insulin alone usually result in weight gain (1–3,27,28).

An increase in the number of lower-limb amputations among patients receiving canagliflozin, as seen in the Canagliflozin Cardiovascular Assessment Study (CANVAS) (29), has prompted speculation that partial inhibition of SGLT1 may interfere with circulation and wound healing. However, the rate of amputation among patients taking sotagliflozin in the inTandem program is lower than both that in CANVAS and the background rate reported for patients with T1D (30). No amputations occurred in inTandem3 (14); the current study and inTandem1 each had a single report of an amputation among sotagliflozin-treated patients.

This trial has some limitations. It was designed to demonstrate HbA_1c reduction in a setting of optimized insulin with more frequent SMBG assessments and insulin adjustments than are typical for clinical practice. Because of study design requirements, HbA_1c and FPG values were masked until the primary end point assessment at week 24. Composite end points are inherently limited, as they give equal weight to all components without regard to frequency or severity; among our composite end points, episodes of DKA, episodes of severe hypoglycemia, and weight gain were rare relative to achievement of HbA_1c <7% (see Supplementary Figs. 4 and 6). In addition, monitoring and patient education mitigated DKA incidence such that it was 0% in the placebo arm. This level of DKA mitigation may not be easily realized in real-world clinical practice.

In phase 3 data published to date, only recently developed automated insulin delivery devices have demonstrated the ability to reduce HbA_1c without increasing the risk for hypoglycemia; other therapeutic interventions with non-SGLT agents have not been successful (10–15,20,31–34). In this study, sotagliflozin significantly reduced HbA_1c while also decreasing documented and severe hypoglycemia events. DKA associated with sotagliflozin occurred at low rates. Sotagliflozin doubled the proportions of patients who achieved target HbA_1c levels without experiencing severe hypoglycemia, DKA, or weight gain. These findings support the use of sotagliflozin in combination with insulin in T1D.

Acknowledgments. The authors thank the inTandem2 trial investigators, staff, and patients for their participation and the following contributors for reviewing the manuscript: John Buse, MD, PhD; Roger Davies, MPhil; Diane Gesty-Palmer, MD, PhD; David Powell, MD; and Kristi Boehm, MS, ELS. Amanda Justice provided medical writing and editorial support, which was funded by Lexicon Pharmaceuticals, Inc. The authors thank Covance Inc. (Princeton, NJ) for operationally executing and performing medical monitoring of this study and Cenduit, LLC (Durham, NC) for providing visualizations of CGM data.

Funding. This study was supported and conducted by Lexicon Pharmaceuticals, Inc. Lexicon Pharmaceuticals, Inc., and Sanofi entered a license agreement effective November 2015 and are collaborating on the development and commercialization of sotagliflozin.

Duality of Interest. T.D. has acted as consultant, advisory board member, steering committee member, or speaker for Abbott Laboratories, Medtronic, Roche, Lexicon Pharmaceuticals, Inc., Menarini Group, Boehringer Ingelheim, AstraZeneca, Novo Nordisk, Sanofi, Dexcom, Inc., and Eli Lilly and Company and has received research grants from Abbott Laboratories, AstraZeneca, Novo Nordisk, Medtronic, and Sanofi. B.C. has received research funding and personal fees from Sanofi and Regeneron Pharmaceuticals, Inc.; research funding from Pfizer; and honoraria from Amgen, AstraZeneca, Genfit SA, Pierre Fabre Group, Eli Lilly and Company, Merck Sharp & Dohme, Novo Nordisk, and Sanofi. P.B., P.L., S.S., and P.S. are employed by Lexicon Pharmaceuticals, Inc., and S.S. holds stock in Lexicon Pharmaceuticals, Inc. M.B. has acted as consultant, advisory board member, or speaker for Boehringer Ingelheim, AstraZeneca, Novo Nordisk, and Merck Sharp & Dohme. H.B. has acted as speaker or advisory board member for Amgen, AstraZeneca, Boehringer Ingelheim, Eli Lilly and Company, Medtronic, Merck Sharp & Dohme, Mylan N.V., Novartis Pharmaceuticals, Novo Nordisk, Pfizer, Sanofi, and Servier Laboratories. E.F. has received lecturer’s fees from AstraZeneca, Bioton, Boehringer Ingelheim, Eli Lilly and Company, Merck Sharp & Dohme, Novo Nordisk, Polfa Tarchomin S.A., Servier Laboratories, and Sanofi and has received consulting fees for serving on advisory boards for Boehringer Ingelheim and Novo Nordisk. J.A.K. serves on the type 1 diabetes steering committee for Lexicon Pharmaceuticals, Inc., and as an advisor for Sanofi and Know Foods. D.K.M. has received consulting fees and fees for serving on a clinical trial executive committee from Boehringer Ingelheim, Sanofi US, Novo Nordisk, and AstraZeneca; consulting fees from Lilly USA LLC; advisory board fees and fees for serving on a clinical trial executive committee from Merck Sharp & Dohme; fees for serving on a data monitoring committee from Janssen Research and Development, LLC, and GlaxoSmithKline; fees for chairing steering committees from Lexicon Pharmaceuticals, Inc.; and fees for serving on a clinical trial executive or steering committee from Eisai Co., Ltd., and Esperion Therapeutics. A.L.P. has participated on advisory boards for Abbott Diabetes Care, Becton, Dickinson and Company, Bigfoot Biomedical, Boehringer Ingelheim, Eli Lilly and Company, Medscape, Merck & Company, Novo Nordisk, Omada Health, and Sanofi; chairs the type 1 diabetes steering committee at Lexicon Pharmaceuticals, Inc.; has consulted for Science 37; has spoken for and received research supplies from Dexcom, Inc.; and has participated in a speaker’s bureau for Novo Nordisk.

Author Contributions. T.D., J.A.K., P.L., D.K.M., A.L.P., S.S., and P.S. conceived and conducted the study and acquired, analyzed, and interpreted data. B.C., M.B., H.B., and E.F. conducted the study and acquired, analyzed, and interpreted data. P.B. contributed to the statistical design, analyzed and interpreted data, and oversaw the statistical analyses conducted by the independent statistician. P.B., P.L., and P.S. approved the protocol and its amendments. P.B., P.L., S.S., and P.S. reviewed the data quality before locking the database. P.L. provided safety oversight for monthly safety reviews. T.D., B.C., P.B., M.B., H.B., E.F., J.A.K., P.L., D.K.M., A.L.P., S.S., and P.S. drafted and critically revised the manuscript. T.D., B.C., P.B., M.B., H.B., E.F., J.A.K., P.L., D.K.M., A.L.P., S.S., and P.S. had full access to the data in the study and had final responsibility for the decision to publish. T.D. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. The inTandem2 24-week results were presented at the European Association for the Study of Diabetes (EASD) Annual Meeting, Lisbon, Portugal, 11–15 September 2017. The inTandem2 52-week results were presented at the 78th Scientific Sessions of the American Diabetes Association in Orlando, FL, 22–26 June 2018.