Using Sotagliflozin In Tandem with Insulin: Weighing the Benefits in Type 1 Diabetes

Written By

Reviewed By


Garg SK, Henry RR, Banks PB, et al. Effects of Sotagliflozin Added to Insulin in Patients with Type 1 Diabetes. N Engl J Med. 2017;377:2337-48.

Patients with type 1 diabetes often have sub-optimal glycemic control.  The gold standard of treatment is basal-bolus insulin or continuous subcutaneous insulin infusion via insulin pump.  However, only a third of patients with type 1 diabetes achieve the American Diabetes Association A1C goal <7%.1  While health professionals have been taught that most patients with type 1 diabetes are lean, obesity is increasingly common and mirrors the rate observed in the general population. Excess weight is associated with insulin resistance and the use of non-insulin agents may be appropriate in patients with type 1 diabetes.2 However, adjunctive agents have failed to demonstrate in clinical trials consistent reductions in hemoglobin A1C (i.e. metformin, DPP4 inhibitors) or have increased the risk of hypoglycemia and diabetic ketoacidosis (i.e. GLP-1 agonists).3 Other than pramlintide, where use is limited by adverse gastrointestinal effects and frequent injections, no other adjunctive agent has been FDA approved to treat type 1 diabetes.


There has been particular interest in using SGLT-2 inhibitors in patients with type 1 diabetes due to their ability to decrease body weight and blood pressure as well as improve glycemic control and cardiovascular outcomes.  A recent meta-analysis which included four randomized, placebo-controlled trials of SGLT-2 inhibitors in patients with type 1 diabetes found significant decreases in A1c, body weight, and insulin dose requirements without increasing the rate of hypoglycemia. However, a significant increase in the incidence of diabetic ketoacidosis (DKA) was observed.4


InTandem3 was a phase III, multicenter, randomized, double-blind, placebo-controlled trial evaluating the safety and efficacy of sotagliflozin, a dual SGLT 1 and 2 inhibitor, in patients with Type 1 diabetes.5 SGLT1 inhibition impairs renal glucose reabsorption, but to a lesser degree than SGLT2. SGLT1 also impairs intestinal glucose absorption and stimulates the release of GLP-1 and polypeptide YY, which suppresses appetite and further improves glycemic control.6


The primary endpoint of this trial was an A1c <7% in the absence of severe hypoglycemia or DKA.  Severe hypoglycemia was not defined by a blood glucose threshold, but rather as an event that required assistance from another person or that led to loss of consciousness or seizure.  Secondary endpoints included change from baseline in body weight, insulin dose, A1c, and systolic blood pressure.


Patients were randomized to receive either sotagliflozin 400 mg (n=699) or placebo (n=703) for 24 weeks administered once daily with the first meal.  Patients had to have type 1 diabetes for at least one year and be on a stable basal insulin dose, not exceeding a 20% difference in total basal insulin dose, for two weeks before the screening visit.  Patients continued their pre-study insulin regimen but the dose could be adjusted during the study. Severe hypoglycemia or DKA during the past month, ≥ 2 episodes of DKA during the past 6 months, or eGFR < 45 mL/1.73m2 were key exclusions for study participation.  Baseline characteristics were similar between groups with an average BMI of 28 kg/m2.  The majority of patients were overweight.  (see Table 1).


Table 1: Baseline Characteristics


Sotagliflozin (N=699)

Placebo (N=703)


43.3 ± 14.2

42.4 ± 14

Female Sex (%)



Duration of Diabetes (years)t

20.5 ± 12.4

19.6 ± 12.1

A1C (%)

8.26 ± 0.96

8.21 ± 0.92

BMI ≥25 (%)



eGFR (mL/min/1.73m2)t

91.51 ± 19.76

92.47 ± 21.89

Daily Total Dose of Insulin (IU/kg)t

0.69 ± 0.28

0.71 ± 0.29

Insulin Dose (IU/day)t



Bolus and corrections


56.88 ± 27.6

29.54 ± 16.29

27.34 ± 16.97


58.35 ± 29.09

29.63 ± 15.54

28.72 ± 19.04

Type of Insulin Therapy (%)t

Subcutaneous injection

Pump (CSII)


424 (60.7)

275 (39.3)


423 (60.2)

280 (39.8)

tmean ± SD

A1c= glycated hemoglobin level, BMI=body mass index, eGFR=estimated glomerular filtration rate, CSII=continuous subcutaneous insulin infusion


A significantly higher percentage of patients in the sotagliflozin arm met the primary endpoint of A1c <7% without severe hypoglycemia or DKA in the modified intent-to-treat population (28.6% vs. 15.2%, respectively, P<0.001).  Significantly greater decreases in A1c were seen in patients with a higher baseline A1c.  A significant decrease in mean total daily insulin was observed (-5.25 unit/day) when compared to placebo. Significant reductions in weight (-2.98 kg) and systolic blood pressure (-3.5 mm Hg) were also observed in the sotogliflozin group.


In terms of safety (see Table 2), more patients in the sotagliflozin group than in the placebo group experienced serious adverse events, leading to 6.3% vs. 3.3% of patients withdrawing from the trial in the sotagliflozin and placebo arms, respectively.  Nausea and diarrhea, consistent with SGLT1 inhibition, were more common with  sotagliflozin use. Genital mycotic infections were also more frequent with sotagliflozin ­— a known, common adverse effect seen with SGLT-2 inhibitors. Patients who used insulin pumps were more likely to have an episode of severe hypoglycemia or DKA when compared to those who used subcutaneous injections.  These events rates are higher than what had been observed in Depict-1 which studied dapagliflozin over 24 weeks in 833 patients with Type 1 diabetes. In Depect-1 there were similar rates of DKA observed in the dapagliflozin and placebo groups.  However, insulin dose reductions were limited to 20% when dapagliflozin was initated.7 


Table 2: Adverse Events [No of patients (%)]






Any adverse event

385 (55.1)

369 (52.5)

Serious adverse events

48 (6.9)

23 (3.3)


Severe hypoglycemia, ≥1 episode

21 (3)

17 (2.4)

Severe hypoglycemia in a patient on  insulin pump, ≥1 episode

10 (3.6)

12 (2.8)

Blood glucose (BG) ≤ 55mg/dL during self-monitoring

528 (75.5)

559 (79.5)


DKA, ≥1 episode

21 (3)

4 (0.6)

DKA in a patient who used insulin pump, ≥1 episode

12 (4.4)

2 (0.7)

Other Adverse Events

Genital mycotic infections

45 (6.4)

15 (2.1)


29 (4.1)

16 (2.3)


14 (2)

9 (1.3)

Bone fracture

4 (0.6)

5 (0.7)


Strengths of this study include the large sample size; multicenter, global study population; and well-matched baseline characteristics.5 There are, however, several limitations. The trial was only 24 weeks, leaving some doubts about the long-term safety and efficacy of this treatment strategy. Investigators were blinded to A1c levels, limiting their ability to make appropriate insulin dose adjustments. The strict criteria used to define severe hypoglycemia likely missed many events, although the self-monitoring blood glucose (SMBG) readings demonstrated higher rates of BG ≤ 55mg/dL in the placebo group.


A rigorous protocol for detecting and treating DKA was employed.  Despite this, 3% of patients on sotagliflozin experienced DKA, including 4.4% of those who used an insulin pump.  Extensive education regarding monitoring for and preventing DKA, such as recognition of DKA signs and symptoms, sick day management, and testing for ketones, is difficult to fit in a typical 20 minute office visit. Moreover, euglycemic DKA has been observed with these agents, making DKA more challenging to detect. If used in the “real world”, it is likely we would see even higher rates of DKA. Considering that this can lead to severe electrolyte disturbances, increased hospital admissions, and life threatening complications, this finding is worrisome.


Studies on cardiovascular outcomes have not been conducted, so it’s unknown if sotogliflozin would provide the same cardiovascular benefits in patients with type 1 diabetes as had been seen with canagliflozin and empagliflozin in patients with type 2. Whether sotogliflozin might increase amputations and bone loss are also unknown. Lastly, patients with eGFR < 45mL/min/1.73m2 were excluded. Many patients with type 1 diabetes have CKD and thus would not be eligible for SGLT-2 inhibitor use.


Continuous glucose monitoring (CGM) was allowed in inTamden3 study but data from this subgroup was not reported.  CGM has the ability to alert patients before they reach dangerously high or low BG readings and it would be interesting to see the incidence of DKA and hypoglycemia in patients who used these devices.  Although patients on insulin pumps had higher rates of DKA, the hybrid-closed loop system was not available at the time of the study. As technology advances toward a true artificial pancreas system, would the use of adjunctive agents affect the pump’s algorithm for insulin dose adjustments?


Despite the increased risk for DKA and severe hypoglycemia, the incidence was relatively low.  One could argue that the improvements in A1C and weight outweigh (no pun intended) the potential harms.  Thus some patients might be good candidates for this therapy such as overweight and obese patients with type 1 diabetes and insulin resistance, hypertension, and difficulty achieving their A1C goal.  On the other hand, patients with recent or frequent hospital admissions for DKA or hypoglycemia, eGFR ≤ 45 mL/min/1.73m2, prone to genital mycotic infections, and infrequent blood glucose monitoring or missing insulin injections should probably avoid sotogliflozin.


Research targeting patients with type 1 diabetes is far less common than clinical trials enrolling patients with type 2 diabetes.  So it’s exciting to see this study and the results.  What do you think about the new SGLT 1 and 2 inhibitor, sotagliflozin? If approved, what do think its role will be in treatment guidelines and your practice?