Sotagliflozin in Combination With Optimized Insulin Therapy in Adults With Type 1 Diabetes: The North American inTandem1 Study

Achieving glycemic targets is difficult for patients with type 1 diabetes (T1D); hypoglycemia, wide glucose fluctuations, and weight gain represent major unresolved challenges (1–6) that threaten patient health (7,8). Mitigating these side effects is essential in T1D management. Pramlintide is approved as an insulin adjunct and may reduce weight as well as HbA_1c, but it requires multiple additional daily injections and is associated with an increased risk of severe hypoglycemia (1,2,9). Agents from other antihyperglycemic classes have not demonstrated sufficient benefit in T1D for manufacturers to pursue market authorization (10–13). Combining insulin with oral sodium–glucose cotransporter (SGLT) inhibitors holds promise to improve glycemic control in T1D without increasing weight or hypoglycemia, and an agent that inhibits SGLT1 further contributes to a blunting and delay of postprandial hyperglycemia (14–16). However, diabetic ketoacidosis (DKA) has emerged as the major clinical concern around SGLT inhibition in T1D (17,18). Possible contributors include decreased basal insulin, a shift to fatty acid oxidation, and increased urinary glucose excretion and volume depletion associated with SGLT2 inhibition (19,20).

Sotagliflozin (LX4211) is a novel dual inhibitor of SGLT1 and SGLT2 that decreases renal glucose reabsorption through systemic SGLT2 inhibition and delays and reduces glucose absorption in the proximal intestine through local SGLT1 inhibition, blunting and delaying postprandial hyperglycemia (21–23). Dual inhibition of SGLT1 and SGLT2 may mitigate DKA risk by reducing glucose absorption in the proximal intestine and thereby diminishing urine glucose excretion and associated water and electrolyte loss; by increasing glucagon-like peptide 1 secretion from the gut, thereby reducing glucagon; and by preserving basal insulin requirements (24–28). We report the efficacy and safety of sotagliflozin combined with insulin delivered as multiple daily injections (MDI) or continuous subcutaneous insulin infusion (CSII) in North American adults with T1D.

This phase 3, multicenter, randomized, double-blind, placebo-controlled, parallel-group study conducted at 75 sites in the U.S. and Canada evaluated the safety and efficacy of oral sotagliflozin 200 or 400 mg combined with insulin in adult patients with inadequately controlled T1D. The study consisted of two double-blind periods: a 24-week treatment period (the primary end point assessment) followed by a 28-week double-blind extension (Fig. 1). Randomization was stratified by insulin delivery method (MDI or CSII), and week −2 HbA_1c values (≤8.5%, >8.5%). A subgroup of patients underwent blinded continuous glucose monitoring (CGM) with a Dexcom G4 monitor (Dexcom, San Diego, CA) during specified 1-week intervals throughout the first 24 weeks.

Starting 6 weeks before randomization, insulin therapy was optimized by adjusting basal and bolus 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 <10 mmol/L (<180 mg/dL). Insulin adjustment continued throughout the trial (Fig. 1).

Institutional review boards approved the protocol and consent forms. All patients provided written informed consent. An independent clinical end point committee, blinded to trial treatment, adjudicated severe hypoglycemia, DKA, major adverse cardiovascular events, drug-induced liver injury, and deaths. An independent insulin dose monitoring committee (IDMC) comprising diabetologists and certified diabetes educators blindly reviewed insulin titration decisions from the start of insulin optimization (6 weeks before baseline) through week 24 to determine whether insulin adjustments were consistent with self-monitoring of blood glucose (SMBG) patterns. An independent data monitoring committee reviewed safety. An independent statistician performed statistical analysis.

The study included men and nonpregnant women aged ≥18 years with T1D using either MDI or CSII for insulin delivery and whose HbA_1c was 7.0–11.0% at screening. Patients with β-hydroxybutyrate (BHB) levels >0.6 mmol/L at screening were excluded. The Supplementary Data lists all inclusion and exclusion criteria.

Patients were randomly assigned 1:1:1 to sotagliflozin 200 mg, sotagliflozin 400 mg, or placebo administered once daily before the first meal of the day. Bolus insulin was reduced by 30% for the first meal after the first dose of study medication on day 1 only (24). Thereafter and throughout the 52-week study, investigators and/or patients adjusted insulin doses according to SMBG data to meet study targets. Based on regulatory feedback, HbA_1c and fasting plasma glucose (FPG) results were masked to study staff from baseline to week 24. After week 12, HbA_1c >11% was unmasked to allow appropriate intervention. During the 28-week extension period, HbA_1c and FPG were unmasked. All patients received urine ketone strips and BHB meters and strips as well as instructions on detecting and treating ketosis (Supplementary Data), urogenital hygiene, and proper hydration. Study centers received recommendations for ketosis and DKA diagnosis and management (Supplementary Data).

The primary end point was the change in HbA_1c from baseline to week 24. The first secondary end point was a composite consisting of the proportion of patients with HbA_1c <7.0% who had no episode of severe hypoglycemia and no episode of DKA at week 24. Other secondary outcomes included the change from baseline to week 24 in body weight, bolus insulin dose, FPG, and scores on the Diabetes Treatment Satisfaction Questionnaire Status (DTSQs) and the 2-item Diabetes Distress Screening Scale (DDS2) (29,30). Additional objectives included the change from baseline in systolic blood pressure (SBP) and diastolic blood pressure at week 12 in all patients and SBP in those with SBP ≥130 mmHg at baseline. HbA_1c, FPG, insulin dose, weight, frequency of documented hypoglycemic events (≤3.9 and ≤3.0 mmol/L [≤70 and ≤55 mg/dL], measured by SMBG), and kidney function were assessed at each study visit throughout the 52-week study. Prespecified composite end points included the proportions of patients meeting HbA_1c targets (<7.0% or ≥0.5% reduction) without experiencing severe hypoglycemia, DKA, or weight gain at 24 and 52 weeks. Safety and tolerability were evaluated throughout the study. DKA diagnosis was based on evidence of metabolic acidosis and other criteria (see Supplementary Data for details).

Efficacy analyses were based on the modified intent-to-treat population, which included all randomized patients who had taken at least one dose of study drug. The primary efficacy end point was the main driver used to calculate the sample size, and the latter was derived by specifying a treatment difference to detect between either dose level of sotagliflozin versus placebo of at least −0.4% and a common SD of this difference as 1.0%. The statistical testing method was planned to control the type 1 family-wise error rate both within and across the primary and secondary end points at α = 0.05. Targeting a statistical power of ≥90% to detect the stated difference and adjusting for dropouts, 250 patients per treatment group was the estimated sample size requirement. 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. For continuous secondary and other efficacy end points, MMRM or ANCOVA was used, with the corresponding end point and baseline value (including first-order interactions in the MMRM) in the model. For binary end points, a Cochran–Mantel–Haenszel test, stratified by the randomization stratification factors, was used. The treatment group comparisons were performed at week 24, with descriptive statistics provided for each clinic visit through week 52. Missing observations at week 24 were imputed as nonresponse. See Supplementary Data for additional details.

Between March 2015 and February 2017, 793 patients were randomly assigned to treatment with placebo (n = 268), sotagliflozin 200 mg (n = 263), or sotagliflozin 400 mg (n = 262), and 218, 228, and 221 completed the study, respectively (Supplementary Fig. 1). Baseline characteristics were similar among groups (Supplementary Table 1). In the total cohort, 473 (59.6%) patients used CSII and 320 (40.4%) used MDI, with similar proportions in each treatment group (Supplementary Table 1).

Insulin optimization over 6 weeks improved HbA_1c by ∼0.65% in the three treatment groups, lowering the mean HbA_1c from ∼8.2% (66 mmol/mol) at screening to 7.54% (58.9 mmol/mol), 7.61% (59.7 mmol/mol), and 7.56% (59.1 mmol/mol), in the placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg groups, respectively, at randomization (baseline). At the week 24 primary end point assessment, the placebo-adjusted least squares mean (LSM) HbA_1c was further reduced by 0.36% (95% CI −0.45 to −0.27) and 0.41% (−0.50 to −0.32) with sotagliflozin 200 and 400 mg, respectively (both P < 0.001 vs. placebo). At 52 weeks, LSM differences from placebo remained significant: −0.25% (−0.37 to −0.14; P < 0.001) and −0.31% (−0.43 to −0.20; P < 0.001) (Fig. 2A and Supplementary Table 2). Approximately 19% of patients from each group had an HbA_1c <7.0% at baseline (Supplementary Table 1). Across the entire study population, 22.8%, 36.9%, and 46.9% of the placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg groups achieved an HbA_1c <7.0% at 24 weeks, and at 52 weeks the proportions were 20.9%, 30.0%, and 35.5%, respectively. Among patients with a baseline HbA_1c ≥7.0% after 6 weeks of insulin optimization, 15.7% of placebo-treated patients, 27.2% of patients receiving sotagliflozin 200 mg (P = 0.003 vs. placebo), and 40.3% of patients receiving sotagliflozin 400 mg (P < 0.001) achieved an HbA_1c <7.0% at week 24. Within this subpopulation after 52 weeks, 18.4%, 27.7%, and 35.4% of patients receiving placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively, had an HbA_1c <7.0%.

Placebo-adjusted differences in FPG were −0.55 mmol/L (−9.8 mg/dL) (95% CI −1.06 to −0.04 mmol/L; P = 0.034) and −0.99 mmol/L (−17.8 mg/dL) (−1.50 to −0.48; P < 0.001) for the 200 and 400 mg doses, respectively, at 24 weeks, and −0.68 mmol/L (−12.2 mg/dL) (−1.28 to −0.08; P = 0.028) and −1.08 mmol/L (−19.4 mg/dL) (−1.69 to −0.47; P < 0.001) at 52 weeks (Supplementary Table 2).

The CGM substudy assessed glycemic variability and included 136 patients (Fig. 3 and Supplementary Table 3). Baseline percent ± SD of time in range (3.9–10.0 mmol/L [70–180 mg/dL]) was 54.19% ± 12.94 and 53.18% ± 13.89 with placebo and sotagliflozin 400 mg, respectively, and time in range increased by 10.4% ± 2.89 (95% CI 4.68 to 16.12; P < 0.001), or 2.50 ± 0.69 h, with sotagliflozin 400 mg relative to placebo. The placebo-adjusted decrease in CGM SD was 0.37 ± 0.17 mmol/L (6.60 ± 3.06 mg/dL) (95% CI −0.70 to −0.30 mmol/L; P = 0.033) with the 400 mg dose.

Total daily insulin dose in the placebo group was within ∼2% of baseline until week 24, after which it increased. In both sotagliflozin groups, after an initial decrease, insulin doses remained stable after week 4 (Fig. 2B). At 24 weeks, the LSM difference from placebo in bolus insulin dose was −5.70% (95% CI −12.82 to 1.42; P = 0.12) with sotagliflozin 200 mg and −12.67% (−19.79 to −5.55; P < 0.001) with sotagliflozin 400 mg. Basal insulin dose increased with placebo while decreasing modestly with sotagliflozin (Supplementary Table 2 and Supplementary Fig. 2A). At 52 weeks, placebo-adjusted percent differences in bolus insulin doses were −5.53% (−14.54 to 3.48; P = 0.23) with sotagliflozin 200 mg and −15.63% (−24.67 to −6.59; P < 0.001) with sotagliflozin 400 mg. Absolute changes in insulin doses appear in Supplementary Table 2 and Supplementary Fig. 2B.

At baseline, mean BMI across all three treatment groups was 29.66 ± 5.387 kg/m², and 44% of participants had a BMI ≥30 kg/m² (Supplementary Table 1). Decreases of 2.35 kg (95% CI −2.85 to −1.85) and 3.14 kg (−3.81 to −2.46) occurred in the sotagliflozin 200 mg group at 24 and 52 weeks, respectively. At week 24, the LSM difference from placebo with sotagliflozin 400 mg was −3.45 kg (−3.95 to −2.94); at week 52 the difference was −4.32 kg (−5.00 to −3.64) (all P < 0.001 vs. placebo) (Fig. 2C and Supplementary Table 2).

At week 12, sotagliflozin 200 and 400 mg were associated with placebo-adjusted mean SBP reductions of 3.5 mmHg (95% CI −5.2 to −1.8; P < 0.001) and 4.2 mmHg (−5.9 to −2.4; P < 0.001), respectively. Diastolic blood pressure also decreased by 1.8 mmHg (−2.8 to −0.7; P = 0.001) and 1.9 mmHg (−2.9 to −0.8; P < 0.001) with sotagliflozin 200 and 400 mg relative to placebo. Among those with SBP ≥130 mmHg at baseline, sotagliflozin 200 and 400 mg reduced SBP by 5.4 mmHg (−9.9 to −1.0; P = 0.017) and 6.6 mmHg (−10.9 to −2.3; P = 0.003) relative to placebo.

In the sotagliflozin 200 and 400 mg arms, estimated glomerular filtration rate initially decreased by 3 and 4 mL/min/1.73 m², respectively, then increased, remaining ∼2 mL/min/1.73 m² below baseline between weeks 24 and 52 (Supplementary Fig. 3 and Supplementary Table 2). The reductions in estimated glomerular filtration rate were reversed 1 week after discontinuation of sotagliflozin (Supplementary Table 2).

At the week 52 visit, mean total cholesterol increased by 0.1 mmol/L in the placebo group and by 0.2 and 0.3 mmol/L in the sotagliflozin 200 and 400 mg groups, respectively. Mean LDL cholesterol increased by 0.1, 0.1, and 0.2 mmol/L with placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. Other lipid changes were small and not considered clinically meaningful.

More patients receiving sotagliflozin than placebo achieved composite end points including an HbA_1c <7.0% plus no severe hypoglycemia, DKA, or weight gain (Supplementary Fig. 4). The proportion achieving an HbA_1c <7.0% without experiencing any severe hypoglycemia or DKA at 24 weeks was 21.64% with placebo, 33.46% with sotagliflozin 200 mg, and 43.51% with sotagliflozin 400 mg (Supplementary Fig. 5). The differences from placebo were statistically significant and remained significant at 52 weeks (sotagliflozin 200 mg: 7.21% [95% CI −0.27 to 14.68; P = 0.049]; sotagliflozin 400 mg: 13.41% [5.67 to 21.15; P < 0.001]). Throughout the study, the difference in proportions of placebo- and sotagliflozin-treated patients who achieved an HbA_1c <7.0% and experienced either severe hypoglycemia (≤3.0%) or DKA (≤1.5%) was not statistically significant (Supplementary Fig. 5). Similar proportions of patients in the sotagliflozin treatment groups achieved an HbA_1c reduction ≥0.5% without severe hypoglycemia or DKA (Supplementary Fig. 5).

The proportion of patients who achieved an HbA_1c <7.0% without weight gain at 24 weeks was 8.58% with placebo, 30.42% with sotagliflozin 200 mg, and 43.51% with sotagliflozin 400 mg (Supplementary Fig. 6 and Supplementary Table 4), with treatment differences from placebo of 21.84% (95% CI 14.97 to 28.71; P < 0.001) and 34.93% (27.68 to 42.18; P < 0.001) for sotagliflozin 200 and 400 mg, respectively. Significantly more patients who achieved an HbA_1c <7.0% lost >5% of body weight while taking sotagliflozin 200 mg (difference from placebo, 5.34% [1.89 to 8.78; P < 0.001]) or 400 mg (12.61% [7.99 to 17.24; P < 0.001]), and few sotagliflozin-treated patients gained weight (Supplementary Fig. 6 and Supplementary Table 4). At 52 weeks, the proportions of patients losing or not gaining weight remained significantly larger in the sotagliflozin groups, and similar patterns were observed among those with an HbA_1c reduction ≥0.5% at 24 and 52 weeks (Supplementary Fig. 6 and Supplementary Table 4).

At 24 weeks, a composite outcome of HbA_1c <7.0%, no weight gain, no severe hypoglycemia, and no DKA was achieved by 7.84%, 27.76%, and 40.84% of patients receiving placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. Differences from placebo were 19.92% (95% CI 13.25 to 26.59; P < 0.001) and 33.00% (25.86 to 40.15; P < 0.001) in the sotagliflozin 200 and 400 mg groups, respectively. Similar treatment differences were maintained at 52 weeks (Supplementary Fig. 6 and Supplementary Table 4).

Treatment satisfaction measured by the DTSQs remained stable in the placebo group while increasing in both sotagliflozin groups. At week 24, the LSM difference between placebo and sotagliflozin 200 mg was 2.5 (95% CI 1.7 to 3.3) and between placebo and sotagliflozin 400 mg was 2.5 (1.8 to 3.3), changes that were statistically significant and clinically meaningful (P < 0.001) (Supplementary Table 2 and Supplementary Fig. 7). Diabetes distress was measured with the DDS2 throughout the study and decreased significantly in both sotagliflozin groups while increasing with placebo (Supplementary Table 2 and Supplementary Fig. 8). At week 24, sotagliflozin-treated patient scores on the DDS2 decreased by 0.7 (−0.9 to −0.4; P < 0.001) and 0.8 (−1.0 to −0.5; P < 0.001) with the 200 and 400 mg doses, respectively, relative to placebo. At week 52, LSM differences from placebo were −0.4 (−0.7 to −0.2; P = 0.003) and −0.5 (−0.8 to −0.2; P < 0.001) for sotagliflozin 200 and 400 mg, respectively.

Over 52 weeks, fewer patients receiving sotagliflozin 200 mg than those receiving placebo reported documented hypoglycemia ≤3.9 mmol/L (≤70 mg/dL) or ≤3.0 mmol/L (≤55 mg/dL), and the frequency of documented hypoglycemia was similar between the placebo and sotagliflozin 400 mg groups (Table 1 and Supplementary Table 2). Documented hypoglycemia ≤3.9 mmol/L occurred at frequencies of 96.1, 84.1, and 90.0 events per person-year in the placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg groups, respectively, with relative rates of 0.88 (95% CI 0.77 to 0.99; P = 0.040) for sotagliflozin 200 mg and 0.93 (0.82 to 1.06; P = 0.28) for sotagliflozin 400 mg versus placebo. The proportion of patients with SMBG ≤3.0 mmol/L from week 51 to 52 was 23.5%, 15.1%, and 13.9% on placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively, with a relative risk for sotagliflozin 200 and 400 mg versus placebo of 0.64 (0.44 to 0.93; P = 0.017) and 0.59 (0.40 to 0.87; P = 0.007), respectively.

Over 52 weeks, 17 (6.5%) patients from each sotagliflozin group and 26 (9.7%) placebo-treated patients experienced ≥1 positively adjudicated severe hypoglycemia event (Table 1). Two patients receiving placebo and one receiving sotagliflozin 200 mg discontinued due to severe hypoglycemia.

Out of 82 patients reporting a serious or nonserious acidosis-related event (Supplementary Table 5), 21 had ≥1 positively adjudicated DKA event: 1 (0.4%; a CSII user) in the placebo group, 9 (3.4%; 8 CSII users) in the sotagliflozin 200 mg group, and 11 (4.2%; 7 CSII users) in the sotagliflozin 400 mg group. In 14 out of 21 DKA events, patients had blood glucose >250 mg/dL (Supplementary Table 5). Positively adjudicated DKA led to discontinuations by four patients in each sotagliflozin group (Table 1).

Mean BHB levels increased by ∼0.1 mmol/L from baseline in both sotagliflozin groups (Supplementary Fig. 9). Over 52 weeks, a BHB ≥0.6 mmol/L occurred in 15.7%, 44.1%, and 45.8% of patients in the placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg groups, respectively.

As shown in Table 1, during 52 weeks of double-blind treatment, the frequency of overall adverse events was similar among the treatment groups: 216 (80.6%), 215 (81.7%), and 209 (79.8%) with placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. The total number of events of special interest occurred at similar rates: 266 (99.3%), 260 (98.9%), and 259 (98.9%) with placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. Sotagliflozin-treated patients reported more serious adverse events (placebo: 20 [7.5%]; 200 mg: 27 [10.3%]; 400 mg: 29 [11.1%]), and severe adverse events occurred in 7 (2.6%) placebo-treated patients and 12 patients from each sotagliflozin group (4.6%). One death due to endocarditis occurred in the placebo group. In the sotagliflozin groups, 13 (4.9%; 200 mg) and 17 (6.5%; 400 mg) patients discontinued due to an adverse event, compared with 11 (4.1%) taking placebo.

Adverse events were mostly mild to moderate in severity. The most common events of special interest were genital mycotic infections (consistent with SGLT2 inhibition) and diarrhea (consistent with SGLT1 inhibition) (Table 1). Among sotagliflozin-treated patients, 24 (9.1%; 200 mg) and 34 (13.0%; 400 mg) reported a genital mycotic infection, compared with 9 (3.4%) receiving placebo. Three patients discontinued due to vulvovaginal events (one receiving placebo and two receiving sotagliflozin 400 mg). Diarrhea occurred in 18 (6.7%), 22 (8.4%), and 27 (10.3%), patients receiving placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively, and led to one discontinuation each in the placebo and sotagliflozin 400 mg groups. Bone fractures occurred in 10 (3.7%), 9 (3.4%), and 5 (1.9%) patients taking placebo, sotagliflozin 200 mg, and sotagliflozin 400 mg, respectively. There was one transmetatarsal amputation that was considered to be not related to the study drug in a patient taking sotagliflozin 400 mg who had a previous history of foot and toe amputations. Four patients (two from each sotagliflozin group) reported malignancies (two breast, one thyroid, and one melanoma); none were considered drug-related.

This study demonstrated the efficacy of sotagliflozin in reducing HbA_1c and improving patient outcomes beyond HbA_1c (31). Sotagliflozin 400 mg was associated with lower insulin doses, FPG, weight, and SBP as well as improved glycemic variability with less hypoglycemia. Sotagliflozin 200 mg did not reduce bolus insulin doses; however, reductions in total insulin dose, FPG, and weight were associated with P values <0.05. The incidence of severe hypoglycemia was lower with sotagliflozin than placebo. DKA was increased with sotagliflozin treatment, yet significantly more people taking sotagliflozin 200 or 400 mg than those receiving placebo had an HbA_1c <7.0% without experiencing severe hypoglycemia, DKA, or weight gain at 24 and 52 weeks. Those receiving sotagliflozin also reported statistically significant improvements in treatment satisfaction and diabetes distress.

The glycemic and weight reductions, as well as composite end points including achievement of HbA_1c targets without severe hypoglycemia, DKA, or weight gain, were comparable to observations from the other inTandem trials. In the 24-week, global inTandem3 study, sotagliflozin 400 mg yielded significant placebo-adjusted reductions in HbA_1c, weight, insulin dose, and blood pressure with reduced hypoglycemia and a low incidence of DKA in patients with T1D (14). In inTandem3, insulin doses were not optimized prior to baseline, a key difference from the current study. The inTandem2 trial was conducted for 52 weeks in Europe with the same design as this study and had similar overall findings, including an even more pronounced improvement in glycemic variability (32). SGLT1 inhibition blunts and delays postprandial glucose absorption. In this and the other inTandem studies, the unique benefit of this mechanism is supported by a lower requirement for bolus insulin, with associated reductions in glycemic variability and hypoglycemia relative to placebo.

Dapagliflozin Evaluation in Patients with Inadequately Controlled Type 1 Diabetes (DEPICT-1) is a phase 3 trial evaluating the use of dapagliflozin in T1D. Baseline HbA_1c was 8.5% (higher than the inTandem1 baseline HbA_1c) after an 8-week lead-in period in which insulin treatment was titrated to individual glucose targets without optimizing it, and placebo-adjusted HbA_1c reductions were 0.4% to 0.5% (15). Patients with HbA_1c <7.5% at baseline were excluded. In the current study, 19% of patients had HbA_1c <7.0% at randomization, and after 6 weeks of insulin optimization, total mean ± SD baseline HbA_1c was 7.6% ± 0.7. The further HbA_1c decrease at 24 weeks was 0.4%. Sotagliflozin 400 mg was also associated with significantly improved measures of glycemic variability, as have been reported with dapagliflozin and canagliflozin (15,33). These improvements were achieved with decreases in total insulin doses that were driven by bolus insulin doses at 24 weeks.

There are multiple pathophysiological mechanisms by which SGLT2 inhibition might increase ketosis, including basal insulin reductions, glucagon increases, and increased urinary glucose excretion, all of which lead to a shift from glucose to fat as source of energy and therefore increase rates of lipolysis in adipose tissue and ketogenesis in the liver, which would increase circulating ketone body levels (19,20). The rate of DKA in this study was higher with sotagliflozin than placebo but still within the range reported in the general T1D population (2). The majority of patients resumed therapy after temporary discontinuation of sotagliflozin due to a DKA event. When using similar adjudication categories, the DKA rates were consistent between DEPICT-1 and this study (34). A contributing factor may have been CSII use, which can increase DKA risk through operational or mechanical failures (35,36). CSII users comprised 60% of the total study population and 75% of patients reporting DKA in the sotagliflozin groups (and the single placebo-treated patient reporting DKA). A preponderance of DKA among CSII users is consistent with other studies of SGLT inhibitors in T1D (14,15). In this and the other inTandem studies, sotagliflozin was associated with mildly increased mean BHB levels of ∼0.1 mmol/L (the smallest increment detectable by point-of-care BHB meters). Increased efforts to mitigate DKA, including protocol revisions, “Dear Dr.” letters, site training, and improved educational materials (patient wallet cards; Supplementary Data), were distributed after U.S. and European regulatory authorities issued safety communications about the risk of DKA with SGLT2 inhibitors in 2015. This enhanced risk mitigation strategy reduced the severity of ketosis for several patients. Further work needs to be done to elucidate the risk factors for DKA and identify the characteristics of patients who would be at higher risk. When SGLT inhibitors are administered, monitoring for ketosis, particularly during metabolically stressful situations, is required. SGLT inhibitors should be discontinued before scheduled surgical procedures, and patients and clinicians should remain in close consultation regarding other forms of behavioral and physiological stress (37).

This study population had a mean baseline BMI of 30 kg/m² and nearly half were obese, consistent with the general adult population with T1D (2,3,38). Sotagliflozin provided a dose-dependent decrease in weight of 3–4 kg (4–5%) relative to placebo, and the magnitude of weight reductions increased throughout the year-long study, demonstrating that sotagliflozin maintained weight loss even in patients receiving intensified insulin therapy.

Sotagliflozin treatment improved diabetes distress and treatment satisfaction in this study and inTandem2 (32). Decreases in weight and/or insulin doses may be linked to these improvements, as suggested by studies reporting improved treatment satisfaction with canagliflozin and pramlintide used as adjunctive therapy for T1D (33,39).

Generalization of these results is limited by protocol details including the population recruited (HbA_1c 7–11%), the supervised insulin optimization efforts during lead-in and for 52 weeks after randomization, more frequent SMBG and ketone testing than in clinical practice, and the expertise and attention of the investigators and study staff. A regulatory request that HbA_1c and FPG be masked until after the primary end point assessment at week 24 also does not reflect clinical practice. The patient education and ketone monitoring called for in the protocol (see details in Supplementary Data) mitigated DKA risk, such that the incidence in the placebo-treated arm was 0.4%. This level of DKA mitigation may be difficult to achieve in clinical practice.

In conclusion, the addition of sotagliflozin to intensified insulin therapy for 1 year in patients with T1D reduced HbA_1c and improved clinical outcomes beyond HbA_1c, including weight reduction, increased time in range, and a lower incidence of severe hypoglycemia. To date, advances in T1D management have been largely limited to insulin formulation and delivery innovations (1,5,6). Although SGLT inhibition is associated with DKA, this risk is manageable with proper education and mitigation plans. The beneficial effects on HbA_1c, glycemic variability, and weight with less hypoglycemia outweigh this risk and support the combination of SGLT inhibitors with insulin as a therapeutic approach for T1D (14–16,24,33,40). After 1 year, substantially greater proportions of patients receiving sotagliflozin versus placebo achieved HbA_1c targets without experiencing severe hypoglycemia, DKA, or weight gain and had improvements in treatment satisfaction. We conclude that sotagliflozin is effective in subjects with T1D and has an acceptable risk-benefit profile.

Acknowledgments. The authors thank the inTandem1 trial investigators, staff, and patients for their participation and the following contributors for reviewing the manuscript: Roger Davies, Diane Gesty-Palmer, David Powell, and Kristi Boehm (Lexicon Pharmaceuticals, Inc., Woodlands, TX). Amanda Justice (independent consultant, Brooklyn, NY) provided medical writing and editorial support, which was funded by Lexicon Pharmaceuticals, Inc. The authors thank Covance, Inc. (Princeton, NJ), for providing the operational execution and medical monitoring of this study and Cenduit, LLC (Durham, NC), for visualization of CGM data.

Funding and Duality of Interest. This study was supported and conducted by Lexicon Pharmaceuticals, Inc. Lexicon and Sanofi entered a license agreement effective November 2015 and are collaborating on the development and commercialization of sotagliflozin. J.B.B. received support from the National Institutes of Health National Center for Advancing Translational Sciences through Grant Award Number UL1TR002489. J.B.B. reports receiving grant support, travel support, and consulting fees paid to his institution from Eli Lilly, GI Dynamics, Merck, AstraZeneca, Sanofi, Intarcia Therapeutics, Lexicon, Orexigen, Takeda, and Novo Nordisk; travel support and consulting fees paid to his institution from Elcelyx Therapeutics, Metavention, vTv Therapeutics, Dance Biopharm, and Adocia; consulting fees paid to his institution from Dexcom, Fractyl, Senseonics, Shenzen HighTide, and NovaTarg; grant support from Medtronic MiniMed, Johnson & Johnson, Boehringer Ingelheim, GlaxoSmithKline, Scion NeuroStim, Theracos, and Bayer; stock options and consulting fees paid to his institution from PhaseBio Pharmaceuticals; stock options from Insulin Algorithms; and travel support and serving on the board of AstraZeneca HealthCare Foundation. S.K.G. reports receiving grant support and travel support from Sanofi, Lexicon, Novo Nordisk, MannKind, Roche Diagnostics, Senseonics, and Medtronic and grant support paid to his institution from Eli Lilly, Dexcom, and Johnson & Johnson. J.Ro. has served on scientific advisory boards and received honoraria or consulting fees from Eli Lilly, Sanofi, Novo Nordisk, Janssen, AstraZeneca, Boehringer Ingelheim, and Intarcia and has received grants/research support from Merck, Pfizer, Sanofi, Novo Nordisk, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Genentech, Janssen, Lexicon, Boehringer Ingelheim, and Intarcia. T.S.B. reports receiving research support from Abbott, Ambra, Ascensia, Becton Dickinson, Boehringer Ingelheim, Calibra, Companion Medical, Dexcom, GlySens, Lexicon, Eli Lilly, MannKind, Medtronic, Novo Nordisk, Sanofi, Senseonics, Versartis, and Xeris; consulting honoraria from AstraZeneca, Ascensia, Becton Dickinson, Calibra, Eli Lilly, Intarcia, Medtronic, Novo Nordisk, and Sanofi; and speaking honoraria from Abbott, Insulet, Eli Lilly, Medtronic, Novo Nordisk, and Sanofi. P.B., P.L., S.S., and P.S. are employed by Lexicon, and S.S. holds stock in Lexicon. B.W.B. reports stock ownership in Aseko and the receipt of consulting fees from Adocia, Intarcia, Janssen, Medtronic, MannKind, Novo Nordisk, and Sanofi and speakers’ bureau fees from AstraZeneca, Lilly/Boehringer Ingelheim, Janssen, Medtronic, MannKind, Novo Nordisk, and Sanofi; in addition, his employer (Atlanta Diabetes Associates) has received grant and research support from Abbott, Becton Dickinson, Dexcom, Diasome, GlaxoSmithKline, Janssen, Lexicon, Lilly/Boehringer Ingelheim, Medtronic, the National Institutes of Health, Novo Nordisk, Sanofi, and Senseonics. T.D. has acted as consultant, advisory board member, or speaker for Abbott, Medtronic, Roche, Lexicon, Menarini, Boehringer Ingelheim, AstraZeneca, Novo Nordisk, Sanofi, Dexcom, and Eli Lilly and has received research grants from Abbott, AstraZeneca, Novo Nordisk, Medtronic, and Sanofi. J.A.K. serves on the type 1 diabetes steering committee for Lexicon and as an advisor for Sanofi and Know Foods. W.S.L. has received honoraria for speaking for Novo Nordisk, Insulet, and Dexcom and for advisory board participation for Novo Nordisk, Insulet, Intarcia, and Sanofi. D.K.M. has received consulting fees and fees for serving on clinical trial executive committees from Boehringer Ingelheim, Sanofi U.S., Novo Nordisk, and AstraZeneca; consulting fees from Lilly USA; 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 and GlaxoSmithKline; fees for chairing an executive committee from Lexicon; and fees for serving on a clinical trial executive or steering committee from Eisai and Esperion. A.L.P. has participated on advisory boards for Abbott Diabetes Care, Becton Dickinson, Bigfoot Biomedical, Boehringer Ingelheim, Eli Lilly, Lexicon, Medscape, Merck, Novo Nordisk, Omada Health, and Sanofi; has consulted for Science 37; has spoken for and received research supplies from Dexcom; and has participated in a speakers’ bureau for Novo Nordisk. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. J.B.B., P.B., T.D., J.A.K., P.L., D.K.M., A.L.P., S.S., and P.S. conceived the study and participated in analysis and interpretation of data. J.B.B., S.K.G., J.Ro., T.S.B., B.W.B., W.S.L., and J.Re. conducted the study, including acquisition, analysis, and interpretation of data. P.B., P.L., and P.S. were involved in approving the protocol and its amendments. P.B., P.L., S.S., and P.S. reviewed the data quality prior to database lock. P.L. provided safety oversight for monthly safety reviews. J.B.B., S.K.G., J.Ro., T.S.B., P.B., B.W.B., T.D., J.A.K., W.S.L., P.L., D.K.M., A.L.P., J.Re., S.S., and P.S. participated in the drafting and critical revision of the manuscript. P.B. contributed to the statistical design, analysis, and interpretation of data and oversaw the statistical analyses conducted by the independent statistician. All authors had full access to the data in the study and had final responsibility for the decision to publish. J.B.B. 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 inTandem1 24-week results were presented at the 77th Scientific Sessions of the American Diabetes Association in San Diego, CA, 9–13 June 2017, and at the 53rd Annual Meeting of the European Association for the Study of Diabetes in Lisbon, Portugal, 11–15 September 2017. The inTandem1 52-week results were presented at the 78th Scientific Sessions of the American Diabetes Association in Orlando, FL, 22–26 June 2018.