28 February 2019

EMA/178275/2019
Committee for Medicinal Products for Human Use (CHMP)

Assessment report
Zynquista

International non-proprietary name: sotagliflozin

Procedure No. EMEA/H/C/004889/0000

Note
Assessment report as adopted by the CHMP with all information of a commercially confidential nature
deleted.

Official address Domenico Scarlattilaan 6 ● 1083 HS Amsterdam ● The Netherlands

Address for visits and deliveries Refer to www.ema.europa.eu/how-to-find-us
Send us a question Go to www.ema.europa.eu/contact Telephone +31 (0)88 781 6000 An agency of the European Union

© European Medicines Agency, 2019. Reproduction is authorised provided the source is acknowledged.
Table of contents
1. Background information on the procedure .............................................. 6
1.1. Submission of the dossier ...................................................................................... 6
1.2. Steps taken for the assessment of the product ......................................................... 7
2. Scientific discussion ................................................................................ 9
2.1. Problem statement ............................................................................................... 9
2.1.1. Disease or condition........................................................................................... 9
2.1.2. Epidemiology .................................................................................................... 9
2.1.3. Aetiology and pathogenesis ................................................................................ 9
2.1.4. Management ..................................................................................................... 9
2.2. Quality aspects .................................................................................................. 10
2.2.1. Introduction .................................................................................................... 10
2.2.2. Active substance ............................................................................................. 10
General information .................................................................................................. 10
Manufacture, characterisation and process controls ....................................................... 11
Specification............................................................................................................. 12
Stability ................................................................................................................... 12
2.2.3. Finished Medicinal Product ................................................................................ 13
Description of the product and Pharmaceutical development .......................................... 13
Manufacture of the product and process controls .......................................................... 15
Product specification ................................................................................................. 15
Stability of the product .............................................................................................. 16
Adventitious agents................................................................................................... 16
2.2.4. Discussion on chemical, pharmaceutical and biological aspects .............................. 16
2.2.5. Conclusions on the chemical, pharmaceutical and biological aspects ...................... 16
2.2.6. Recommendation(s) for future quality development ............................................. 16
2.3. Non-clinical aspects ............................................................................................ 17
2.3.1. Pharmacology ................................................................................................. 17
2.3.2. Pharmacokinetics............................................................................................. 17
2.3.3. Toxicology ...................................................................................................... 18
2.3.4. Ecotoxicity/environmental risk assessment ......................................................... 20
2.3.5. Discussion on non-clinical aspects...................................................................... 20
2.3.6. Conclusion on non-clinical aspects ..................................................................... 22
2.4. Clinical aspects .................................................................................................. 22
2.4.1. Introduction .................................................................................................... 22
2.4.2. Pharmacokinetics............................................................................................. 25
2.4.3. Pharmacodynamics .......................................................................................... 28
Relationship between plasma concentration and effect................................................... 33
2.4.4. Discussion on clinical pharmacology ................................................................... 36
2.4.5. Conclusions on clinical pharmacology ................................................................. 40
2.5. Clinical efficacy .................................................................................................. 40
2.5.1. Dose response studies...................................................................................... 40
2.5.2. Main studies ................................................................................................... 44
Clinical studies in special populations .......................................................................... 78
Analysis performed across trials (pooled analyses) ........................................................ 78

Assessment report
EMA/178275/2019 Page 2/125
Supportive study ...................................................................................................... 80
2.5.3. Discussion on clinical efficacy ............................................................................ 81
2.5.4. Conclusions on clinical efficacy .......................................................................... 87
2.6. Clinical safety .................................................................................................... 88
Patient exposure ....................................................................................................... 88
Adverse events ......................................................................................................... 89
Serious adverse events and deaths ............................................................................. 99
Laboratory findings ................................................................................................... 99
Safety in special populations .................................................................................... 100
Immunological events ............................................................................................. 100
Safety related to drug-drug interactions and other interactions ..................................... 101
Discontinuation due to AEs ....................................................................................... 101
2.6.1. Discussion on clinical safety ............................................................................ 101
2.6.2. Conclusions on clinical safety .......................................................................... 107
2.7. Risk Management Plan ...................................................................................... 108
2.8. Pharmacovigilance ............................................................................................ 111
2.9. New Active Substance ....................................................................................... 111
2.10. Product information ........................................................................................ 112
2.10.1. User consultation ......................................................................................... 112
2.10.2. Quick Response (QR) code ............................................................................ 112
2.10.3. Additional monitoring ................................................................................... 112
3. Benefit-Risk Balance............................................................................ 112
3.1. Therapeutic Context ......................................................................................... 112
3.1.1. Disease or condition....................................................................................... 112
3.1.2. Available therapies and unmet medical need ..................................................... 112
3.1.3. Main clinical studies ....................................................................................... 113
3.2. Favourable effects ............................................................................................ 113
3.3. Uncertainties and limitations about favourable effects ........................................... 115
3.4. Unfavourable effects ......................................................................................... 116
3.5. Uncertainties and limitations about unfavourable effects ....................................... 117
3.6. Effects Table .................................................................................................... 118
3.7. Benefit-risk assessment and discussion ............................................................... 120
3.7.1. Importance of favourable and unfavourable effects ............................................ 120
3.7.2. Balance of benefits and risks ........................................................................... 122
3.7.3. Additional considerations on the benefit-risk balance ......................................... 122
3.8. Conclusions ..................................................................................................... 122
4. Recommendations ............................................................................... 122

Assessment report
EMA/178275/2019 Page 3/125
List of abbreviations
A1C haemoglobin A1C
AUC area under the curve
BCS Biopharmaceutics Classification system
bid twice daily
CGM continuous glucose monitoring
CHMP Committee for Medicinal Products for Human use
CFU Colony Forming Units
CPP Critical process parameter
CQA Critical Quality Attribute
CTD Common technical document
CYP450 cytochrome P450
DBP diastolic blood pressure
DEXA dual-energy X-ray absorptiometry
DDI drug-drug-interaction
DKA diabetic ketoacidosis
DoE Design of experiments
DSC Differential Scanning Calorimetry
DTSQ Diabetes Treatment Satisfaction Questionnaire
EC European Commission
eGFR estimated glomerular filtration rate
EU European Union
FMEA Failure mode effects analysis
FPG fasting plasma glucose
fu fraction unbound
GC Gas Chromatography
GI gastrointestinal
GLP-1 glucagon-like peptide 1
GMP Good Manufacturing Practice
HDPE High Density Polyethylene
HI hepatic impairment
HPLC High performance liquid chromatography
ICH International Conference on Harmonisation of Technical Requirements for Registration of
Pharmaceuticals for Human Use
IPC In-process control
IR Infrared
ITT intent-to-treat
KF Karl Fischer titration
LTE long-term extension
MA Marketing Authorisation
MAG mean absolute glucose
MAGE mean amplitude of glucose excursion
MAR missing at random
MNAR missing not at random
MI multiple imputation
mITT modified Intention to Treat
MMRM mixed model for repeated measures
MS Mass Spectrometry

Assessment report
EMA/178275/2019 Page 4/125
NMR Nuclear Magnetic Resonance
NOR Normal Operating Range
OGTT oral glucose tolerance test
PAR Proven Acceptable Range
PD pharmacodynamic(s)
Ph. Eur. European Pharmacopoeia
PK pharmacokinetic(s)
PMM patten mixture model
PopPK population PK analysis
PP per protocol
PPG postprandial glucose
PSDD patient satiety daily diary
PSDD-NRS patient satiety daily diary numeric rating scale
PVC/PCTFE/PVC-Al polyvinyl chloride / polychlorotrifluoroethylene / polyvinyl chloride-aluminium
PYY peptide YY
QbD Quality by design
qd once daily
QTPP Quality target product profile
RI renal impairment
SBP systolic blood pressure
SGLT1 sodium glucose co-transporter 1
SGLT2 sodium glucose co-transporter 2
SH severe hypoglycaemia
SMBG self-monitoring of blood glucose
SmPC Summary of Product Characteristics
T1DM type 1 diabetes mellitus
T2DM type 2 diabetes mellitus
TAMC Total Aerobic Microbial Count
TGA Thermo-Gravimetric Analysis
TTC Threshold of toxicological concern
TYMC Total Combined Yeasts/Moulds Count
UGT UDP-glucuronosyltransferase
USP United States Pharmacopoeia
USP/NF United States Pharmacopoeia/National Formulary
UV Ultraviolet
XRPD X-Ray Powder Diffraction

Assessment report
EMA/178275/2019 Page 5/125
1. Background information on the procedure

1.1. Submission of the dossier

The applicant sanofi-aventis groupe submitted on 7 March 2018 an application for marketing
authorisation to the European Medicines Agency (EMA) for Zynquista, through the centralised procedure
falling within the Article 3(1) and point 3 of Annex of Regulation (EC) No 726/2004. The eligibility to the
centralised procedure was agreed upon by the EMA/CHMP on 20 July 2017.

The applicant applied for the following indication: Zynquista is indicated as an adjunct to insulin therapy
to improve glycaemic control in adults with type 1 diabetes mellitus.

The legal basis for this application refers to:

Article 8.3 of Directive 2001/83/EC - complete and independent application

The application submitted is composed of administrative information, complete quality data, non-clinical
and clinical data based on applicants’ own tests and studies and/or bibliographic literature
substituting/supporting certain test(s) or study(ies).

Information on Paediatric requirements

Pursuant to Article 7 of Regulation (EC) No 1901/2006, the application included an EMA Decision(s)
P/0337/2017 on the agreement of a paediatric investigation plan (PIP).

At the time of submission of the application, the PIP P/0337/2017 was not yet completed as some
measures were deferred.

Information relating to orphan market exclusivity

Similarity

Pursuant to Article 8 of Regulation (EC) No. 141/2000 and Article 3 of Commission Regulation (EC) No
847/2000, the applicant did not submit a critical report addressing the possible similarity with authorised
orphan medicinal products because there is no authorised orphan medicinal product for a condition
related to the proposed indication.

New active Substance status

The applicant requested the active substance sotagliflozin contained in the above medicinal product to be
considered as a new active substance, as the applicant claims that it is not a constituent of a medicinal
product previously authorised within the European Union.

Scientific advice

The applicant received Scientific advice from the CHMP:

Scientific advice date Area

EMEA/H/SA/2417/1/1012/III 5 November 2012 non-clinical, clinical

EMEA/H/SA/2417/2/2014/III 24 July 2014 non-clinical, clinical

Assessment report
EMA/178275/2019 Page 6/125
 Scientific advice date Area

EMEA/H/SA/2417/2/FU/1/2015/II 19 November 2015 clinical

EMEA/H/SA/2417/3/2017/I 23 February 2017 quality

The Scientific Advice pertained to the following quality, non-clinical and clinical aspects of the dossier:

• Demonstration of Quality comparability of drug substance and drug product manufactured by

different suppliers and adequacy of stability data plans
• Adequacy of non-clinical studies investigating bone safety, juvenile toxicity and potential toxicity
of metabolites
• Appropriateness of planned clinical food-effect studies
• Appropriateness of a clinical phase 2b study design investigating adolescents and young adults
with T1DM, in particular: definition of study population, sample size, treatment duration, efficacy
endpoints, dose finding
• Appropriateness of the design of two proposed pivotal phase 3 clinical studies, in particular:
definition of study population, screening and run-in periods, insulin-treatment management,
dose and dose regimen, sample size, primary and secondary efficacy endpoints, safety database
at the time of MAA, evaluation of severe hypoglycaemia risk and diabetic ketoacidosis:
monitoring, risk evaluation and management
• Appropriateness of plans to assess cardiovascular safety

1.2. Steps taken for the assessment of the product

The Rapporteur and Co-Rapporteur appointed by the CHMP were:

Rapporteur: Kristina Dunder Co-Rapporteur: Bart Van der Schueren

The application was received by the EMA on 7 March 2018

The procedure started on 29 March 2018

The Rapporteur's first Assessment Report was circulated to all CHMP 18 June 2018
members on

The Co-Rapporteur's first Assessment Report was circulated to all CHMP 18 June 2018
members on

The PRAC Rapporteur's first Assessment Report was circulated to all PRAC 2 July 2018
members on

The CHMP agreed on the consolidated List of Questions to be sent to the 26 July 2018
applicant during the meeting on

The applicant submitted the responses to the CHMP consolidated List of 14 Sept 2018
Questions on

The Rapporteurs circulated the Joint Assessment Report on the responses 22 Oct 2018
to the List of Questions to all CHMP members on

The PRAC agreed on the PRAC Assessment Overview and Advice to CHMP 31 Oct 2018
during the meeting on

Assessment report
EMA/178275/2019 Page 7/125
 The CHMP agreed on a 1st List of Outstanding Issues to be addressed in 15 Nov 2018
writing and/or in an oral explanation to be sent to the applicant on

Ad Hoc Expert group was convened to address questions raised by the 21 Nov 2018
CHMP on

The CHMP considered the views of the Ad Hoc Expert group as presented
in the minutes of this meeting.

The applicant submitted the responses to the CHMP List of Outstanding 03 Jan 2019
Issues on

The Rapporteurs circulated the Joint Assessment Report on the responses 16 Jan 2019
to the List of Outstanding Issues to all CHMP members on

The outstanding issues were addressed by the applicant during an oral 29 Jan 2019
explanation before the CHMP during the meeting on

The CHMP agreed on a 2nd List of Outstanding Issues to be addressed in 31 Jan 2019
writing and/or in an oral explanation to be sent to the applicant on

The applicant submitted the responses to the CHMP List of Outstanding 05 Feb 2019
Issues on

The Rapporteurs circulated the Joint Assessment Report on the responses 13 Feb 2019
to the List of Outstanding Issues to all CHMP members on

The CHMP, in the light of the overall data submitted and the scientific 28 Feb 2019
discussion within the Committee, issued a positive opinion for granting a
marketing authorisation to Zynquista on

Assessment report
EMA/178275/2019 Page 8/125
2. Scientific discussion

2.1. Problem statement

2.1.1. Disease or condition

The indication proposed for Zynquista is:

“Zynquista is indicated as an adjunct to insulin therapy to improve glycaemic control in adults with type
1 diabetes mellitus with a Body Mass Index (BMI) ≥ 27 kg/m2, who have failed to achieve adequate
glycaemic control despite optimal insulin therapy.”

2.1.2. Epidemiology

The worldwide incidence of T1DM varies by at least 100-fold, being highest in Finland and Sardinia (~40
to 60/100 000) and lowest in Venezuela and China (<1/100 000). The incidence has been increasing
worldwide at an annual rate of approximately 3% and T1DM currently affects approximately 30 million
adults worldwide. On average, 40 000 people are diagnosed with T1DM each year in the United States.

2.1.3. Aetiology and pathogenesis

Type 1 diabetes mellitus is an autoimmune disease characterized by rapidly progressive pancreatic β-cell
destruction leading to a state of absolute insulin deficiency. Exogenous insulin is required for survival and
treatment in T1DM. Lack of insulin results in diabetic ketoacidosis (DKA), a condition which is potentially
life-threatening. About 3% of patients with T1DM initially presents with DKA.

2.1.4. Management

Therapeutic options for type 1 diabetes are limited to intensive insulin or insulin analogue therapy
delivered subcutaneously by MDI or insulin pump. The only non-insulin therapy indicated as adjunct to
insulin treatment of T1DM is Symlin (pramlintide acetate injection), approved in 2005 in the US. Other
non-insulin therapies for T1DM have been pursued over the past 2 decades, but no clear benefit has yet
been shown. Segmental pancreatic and islet cell transplantation continues to be explored but to date are
limited to very few patients.

Based largely on the results of the DCCT, current treatment guidelines recommend that patients with
T1DM be treated to a goal HbA1c of <7.0% with intensive insulin therapy. Despite significant advances in
insulin therapies, delivery methods and management, 70% of T1DM patients do not achieve optimal
glycaemic control and are at increased risk of microvascular complications caused by chronic
hyperglycaemia. Furthermore, some patients with T1DM experience wide fluctuation in their level of
glucose control on currently available insulin regimens and high glucose variability may affect quality of
life.

Intensive insulin therapy is associated with an increased risk of hypoglycaemia, which can be
life-threatening and may prevent patients from achieving optimal glycaemic control. Furthermore,
intensive insulin therapy is also associated with excessive weight gain and with peripheral insulin
resistance, both being risk factors for hypertension, and CV disease. Approximately 60% of adults living
with T1DM are overweight or obese, 60% have dyslipidaemia and 40% have hypertension, and are
therefore at risk of developing macrovascular complications.

During insulin therapy, DKA may occur in situations where insulin administration is interrupted or when
the insulin need is increased, e.g. with severe infections. In the literature, the incidence of DKA in T1DM

Assessment report
EMA/178275/2019 Page 9/125
patients shows large variations with a range from 0-56 per 1000 person-years. The prevalence of DKA
decreases with increasing age.

The fact that most patients fail to achieve glycaemic targets and are at risk of hypoglycaemia and
excessive weight gain on current type 1 treatments constitute an unmet need that could potentially be
addressed by a new treatment used as an adjunct to insulin.

About the product

Sotagliflozin is an orally delivered, small molecule dual inhibitor of SGLT1 and SGLT2. Sotagliflozin
improves glycaemic and metabolic control through dual inhibition – local inhibition of SGLT1 in the gut
and systemic SGLT2 inhibition in the proximal renal tubule. Inhibition of SGLT1 delays and reduces
glucose absorption in the proximal intestine, resulting in a blunting and delay of postprandial
hyperglycaemia. Inhibition of SGLT2 reduces renal glucose reabsorption and lowers the renal threshold
for glucose, and thereby increases urinary glucose excretion.

Sotagliflozin is administered by the oral route. The proposed commercial formulation for sotagliflozin is a
200 mg immediate release film-coated tablet. The recommended dose is 200 mg sotagliflozin once daily
(administered as 1 tablet of 200 mg). The dose may be increased to 400 mg once daily (administered as
2 tablets of 200 mg).

Type of Application and aspects on development

The development program has in all essential aspects followed the EMA Guideline “Clinical investigation of
medicinal products in the treatment or prevention of diabetes mellitus” (CHMP/EWP/1080/00 Rev. 1) and
the scientific advice given (see sections 1.1 and 2.5.3 of this report).

2.2. Quality aspects

2.2.1. Introduction

The finished product is presented as film-coated tablets containing 200 mg of sotagliflozin as active
substance.

The other ingredients are: microcrystalline cellulose, croscarmellose sodium, colloidal anhydrous silica,
magnesium stearate, talc, a film-coating agent (polyvinyl alcohol, macrogol 4000, titanium dioxide
(E171), talc, indigo carmine aluminium lake (E132)) and black ink (shellac, black iron oxide (E172),
propylene glycol).

The product is available in polyvinyl chloride/polychlorotrifluoroethylene/ polyvinyl chloride/aluminium

(PVC/PCTFE/PVC-Al) blisters as described in section 6.5 of the SmPC.

2.2.2. Active substance

General information

The chemical name of sotagliflozin is (5S)-Methyl

5-C-[4-chloro-3-[(4-ethoxyphenyl)methyl]phenyl]-1-thio-β-L-xylopyranoside corresponding to the
molecular formula C21H25ClO5S. It has a relative molecular mass of 424.94 and the following structure:

Assessment report
EMA/178275/2019 Page 10/125
 Figure 1: active substance structure

The active substance is a white to off-white non-hygroscopic solid that shows low aqueous solubility
independent of pH.

The chemical structure of sotagliflozin was elucidated by a combination of spectroscopic methods

[infrared spectroscopy (IR), nuclear magnetic resonance spectroscopy (NMR) (comprising 1D and 2D
experiments, mass spectrometry (MS), ultraviolet absorption spectroscopy (UV)] and elemental analysis.

The solid state properties of the active substance were measured by X-ray powder diffraction (XRPD),
Differential scanning calorimetry (DSC), Thermogravimetric analysis (TGA). A number of polymorphic
forms of sotagliflozin have been identified. It has been conclusively demonstrated via XRPD pattern and
DSC thermogram that the proposed manufacturing process consistently produces the same crystalline
form of the active substance and that this form does not change upon storage.

Sotagliflozin exhibits stereoisomerism due to the presence of five chiral centres. The molecule contains
five stereogenic centres and 32 stereoisomers are theoretically possible. However, sotagliflozin is
produced via a highly stereoselective manufacturing process which has been demonstrated to
consistently result in a single stereoisomer. The assigned absolute stereochemistry (2S,3R,4R,5S,6R) of
sotagliflozin at its stereocentres has been confirmed by single crystal X-ray diffraction analysis.

Manufacture, characterisation and process controls

Sotagliflozin is synthesised in five main steps from two well defined starting materials with acceptable
specifications.

The manufacturing process is operated by a traditional approach using target set points/normal operating
ranges (NORs) for the applied process parameters. Process parameter studies on the manufacturing
process are described and proven acceptable ranges (PARs) have been established for each process step
based on univariate experiments but supportive data are not included in the documentation. This is
considered acceptable as the PARs presented in the file constitute manufacturing process development
knowledge and have not been included in the proposed manufacturing process for regulatory flexibility.

Based on the risk assessment and development work, no critical parameters (CPPs) have been identified
at the different stages of the manufacturing process.

The proposed GMP starting materials are considered acceptable based on requirements of the ICH Q11
guideline and taking account of Scientific Advice provided by CHMP prior to submission of this MA
application. Satisfactory information with regard to suppliers, route of synthesis, specifications, analytical
methods and batch analysis has been provided for the starting materials. The specifications presented for
the reagents, solvents and auxiliary materials used in the process are considered adequate.

Adequate in-process controls are applied during the synthesis. The specifications and control methods for
intermediate products, starting materials and reagents have been presented and are considered
acceptable.

The characterisation of the active substance and its impurities are in accordance with the EU guideline on
chemistry of new active substances. Potential and actual impurities were well discussed with regards to
their origin and characterised. The applicant has presented a comprehensive understanding of impurities

Assessment report
EMA/178275/2019 Page 11/125
generated during each step and their fate through the process. Possible impurities originating from the
starting materials and intermediates, and their fate and purge during the downstream process to the final
active substance, have overall been satisfactorily discussed and documented based on spiking studies
and batch analysis data. The proposed impurities specifications for the starting materials, intermediates
and the final active substance are deemed justified through these investigations.

The stereochemical purity of sotagliflozin has been demonstrated and is assured by an appropriate control
strategy.

Mutagenic impurities have been addressed in line with the requirements of the ICH M7 guideline. An
appropriate control strategy for mutagenic impurities is in place. It is informed that no Class 1 organic
solvents are used in the active substance manufacturing process. An appropriate control strategy for
residual Class 2 and 3 solvents used in the active substance manufacturing process is in place.

No Class 1, 2 or 3 metals are used in the active substance manufacturing process. An appropriate control
strategy for relevant metals is in place.

The commercial manufacturing process for the active substance was developed in parallel with the clinical
development program. The process development history is described in sufficient detail. Over time,
changes have been made to the process in order to improve manufacturing efficiency while ensuring
quality. These changes are related to the solvents and reagents used. The overall synthetic route has not
changed. Satisfactory comparative analytical data for batches manufactured with the former and current
process are provided. The quality of the active substance used in the various phases of the development
is considered to be comparable with that produced by the proposed commercial process.

The active substance packaging complies with the EC directive EC 10/2011 as amended.

Specification

The active substance specification includes tests for appearance, identification (IR, HPLC), assay (HPLC),
related substances (HPLC), water content (KF), residual solvents (GC), sulfated ash (Ph. Eur.),
polymorphic form (DSC), particle size distribution (Laser diffraction) and microbiological examination (Ph.
Eur.).

All appropriate test parameters are included in the active substance specification and the corresponding
acceptance limits are considered adequate and well justified. The analytical methods used have been
adequately described and (non-compendial methods) appropriately validated in accordance with the ICH
guidelines. Satisfactory information regarding the reference standards used for assay and related
substances testing has been presented.

Batch analysis data of the active substance are provided covering the proposed commercial batch size
range. The results are within the specifications and consistent from batch to batch. The provided batch
analysis data confirm the consistency of the manufacturing process and the quality of the active
substance.

Stability

Stability studies on three batches (two production and one pilot scale) of the active substance stored in
the proposed commercial packaging for up to 36 months under long term conditions (25 °C/60% RH) and
up to 12 months under accelerated conditions (40 ºC/75% RH) have been provided. The following
relevant stability indicating parameters were tested: appearance, identity, assay, impurities, water
content, particle size, crystal form and microbiological quality. The analytical methods used were the
same as for release and were stability indicating.

Assessment report
EMA/178275/2019 Page 12/125
All tested parameters were within the specifications and no meaningful changes or trends in any of the
monitored parameters have been observed during the evaluated storage periods.

Photostability testing following the ICH guideline Q1B was performed on one batch. A slight change in
appearance, without any quantifiable degradation, was observed in the photostability study. Based on
these results, sotagliflozin is considered to be sensitive to light and must therefore be stored in a
secondary container to provide adequate light protection.

A stress testing study has been conducted on sotagliflozin active substance in order to demonstrate the
stability indicating nature of the HPLC assay/purity method and to study the degradation pathway. The
active substance remained stable following subjection to the heat/humidity and photolytic stress
conditions in the solid state while degradation was observed during the acidic, alkaline and oxidative
stress conditions in the solution state. The degradation impurities that appeared during the stress testing
in the solution state are separated from the sotagliflozin peak, demonstrating that the method is suitably
stability indicating.

The stability results indicate that the active substance manufactured by the proposed supplier is
sufficiently stable. The stability results justify the proposed retest period of 48 months without any
specific storage condition when packaged in the proposed container as described in Section S.6.

2.2.3. Finished Medicinal Product

Description of the product and Pharmaceutical development

The finished product is blue oval-shaped film-coated tablets printed with “2456” on one side in black ink
(tablet length: 14.2 mm, tablet width: 8.7 mm) containing 200 mg of sotagliflozin.

All excipients are well known pharmaceutical ingredients and their quality is compliant with Ph. Eur
standards. There are no novel excipients used in the finished product formulation. The list of excipients is
included in section 6.1 of the SmPC and in paragraph 2.2.1 of this report.

Tests for relevant functionality-related characteristics are included in the excipient specifications. A
ready-to-use film-coating agent is used to allow ease of swallowing of the tablet and black ink is utilised
for imprint. Purified water is used for the film-coating suspension and is removed in the manufacturing
process.

Pharmaceutical Development

The finished product has been developed by a traditional approach according to ICH Q8 and no regulatory
flexibility is applied for. The applicant has however made use of a mixture prior knowledge and quality by
design (QbD) approaches during development. The quality target product profile (QTPP) was defined as
summarised in Table 1.
Table 1 Quality target product profile for sotagliflozin 200 mg film-coated tablets

QTTP element Target

Therapeutic indication Adult patients with diabetes mellitus

Drug delivery requirement Oral immediate release

Dosage form and strength Film coated tablet 200 mg

Appearance of dosage form Blue oval shaped, imprinted on one side

Dose regimen Once a day

Assessment report
EMA/178275/2019 Page 13/125
 Primary packaging Blister (for US & EU), HDPE bottles (for US)

Shelf-life and storage conditions At least 36 months at 2-30°C

The formulation and manufacturing development have been evaluated through the use of prior
knowledge, risk assessments and design of experiments to identify the critical product quality attributes
and critical process parameters. A risk analysis was performed using the failure mode effect analysis
(FMEA) method in order to define critical process steps and process parameters that may have an
influence on the finished product quality attributes. The risk identification was based on the prior
knowledge of products with similar formulations and manufacturing processes as well as on the
experience from formulation development, process design and scale-up studies. The critical quality
attributes (CQAs) and potential critical process parameters (CPPs) have been adequately investigated.
The resulting formulation and manufacturing process are considered well justified.

Formulation development

Sotagliflozin is a white to off-white powder and its crystalline form is used for the film-coated tablets. The
solubility of sotagliflozin in water is very low and the solubility is pH independent. Permeability studies
have been conducted results of which suggest good permeability and the applicant classifies sotagliflozin
as a BCS Class 2 compound (low solubility, good permeability).

The proposed particle size specification has been set on the basis of development data generated on a
large number of active substance batches used for manufacture of clinical development product batches
and by taking the possible impact on dissolution performance of the finished product as well as the
analytical variability of the particle size test method into consideration. The formulation development has
been described in detail.

For the commercial tablet, the critical quality attributes identified were assay and identity of sotagliflozin,
uniformity of dosage units, dissolution, related substances, moisture and microbiological quality. A blue
film-coating was chosen, and black imprinting on one side of the tablet was added. All (historic) changes
made to the formulations have been described and accounted for in detail in the dossier.

The dissolution method development has been described in sufficient detail and the final selected method
apparatus and parameters are considered justified. The proposed limit at release and during shelf-life has
been derived from dissolution profiles obtained from clinical development and registration batches of the
finished product. The same dissolution method has been utilised throughout development and the
discriminatory power of the dissolution method has been adequately demonstrated by testing tablets
manufactured with varied manufacturing parameters and excipient contents.

Manufacturing process development

The manufacturing process is a standard process comprising mixing, granulation, lubrication,

compression, film-coating and printing. The process development has been described in detail.

An acceptable and justified control strategy has been presented. The process parameters and process
controls have satisfactorily been derived from adequate prior knowledge and suitable design of
experiments. The developed process is adequately reflected and described with normal operating ranges
(NORs) and in-process controls (IPCs). The critical steps, process parameters and process controls are
adequately described.

Packaging development

The primary packaging is white opaque standard polyvinyl chloride / polychlorotrifluoroethylene /

polyvinyl chloride-aluminium (PVC/PCTFE/PVC-Al) blisters. The secondary package is a cardboard box.

Assessment report
EMA/178275/2019 Page 14/125
The container closure system has been acceptably described and the control specifications for the
packaging components are considered adequate. The material complies with Ph. Eur. and EC
requirements. The choice of the container closure system has been validated by stability data and is
adequate for the intended use of the product.

Manufacture of the product and process controls

The manufacturing process consists of nine main steps: pre-blending, granulation, blending, final
blending, compression, film-coating suspension, film-coating, printing and packaging. The process is
considered to be a standard manufacturing process.

The manufacturing process has been sufficiently well described in a narrative and flow charts. The
applicant has acquired substantial knowledge of the process during development. The process
parameters and in-process controls have been adequately derived from the manufacturing process
development and are adequate for this type of manufacturing process / pharmaceutical form.

The critical steps of the process have been identified. The proposed process parameters and in-process
controls for the critical steps are well and are considered justified.

Bulk stability data, along with an evaluation of the transport conditions, have been provided for the
coated tablets prior to packaging in the intended commercial packaging and the proposed maximum
holding time of 12 months is acceptable. It has been confirmed that the finished product shelf-life is
calculated in accordance with the guideline ‘Note for guidance on start of shelf-life of the finished dosage
form’ (CPMP/QWP/072/96).

The process has been developed and scaled up at the proposed manufacturing site. The process has been
studied in detail at pilot and commercial scale using the same type of equipment. A pre-qualification batch
at production scale has also been manufactured. It has been demonstrated that the manufacturing
process is capable of producing the finished product of intended quality in a reproducible manner. The
process will be formally validated with three consecutive production scale batches prior to marketing. An
acceptable process validation scheme has been provided.

Product specification

The finished product specifications includes appropriate tests for this kind of dosage form; appearance,
identification (HPLC, UV), assay (HPLC), degradation products (HPLC), dissolution (Ph. Eur.), uniformity
of dosage units (Ph. Eur.), water content (KF) and microbiological examination (Ph. Eur.).

The finished product is released on the market based on the above release specifications, through
traditional final product release testing. The specifications are considered justified based on batch
analysis, stability data and regulatory guidelines / requirements. Based on the risk analysis and data
presented for elemental impurities in the finished product as per the ICH Q3D guideline, it can be
concluded that it is not necessary to include any elemental impurity controls in the finished product
specification.

The analytical methods used have been adequately described and appropriately validated in accordance
with the ICH guidelines. Satisfactory information regarding the reference standards used for assay and
degradation products testing has been presented.

Batch analysis results are provided for a number of batches (10 in total) manufactured for Phase 3 clinical
trials, for development studies, for primary stability studies and batches representing the commercial
scale and manufacture. The results for all reported batches comply with the acceptance criteria
confirming the consistency of the manufacturing process and its ability to manufacture to the intended
product specification.

Assessment report
EMA/178275/2019 Page 15/125
Stability of the product

Stability data from three pilot scale batches of finished product stored for up to 24 months under long
term conditions (25 ºC / 60% RH) and for up to 6 months under accelerated conditions (40 ºC / 75% RH)
according to the ICH guidelines were provided.

The batches of finished product are representative of those proposed for marketing and were packed in in
different primary container closure systems: blisters and HDPE bottles.
The stability studies of the primary batches are conducted according to a reduced matrix design with
these different packages, which was considered acceptable.

The batches were tested for appearance, assay, degradation products, dissolution, water content and
microbiological quality. The analytical procedures used were stability indicating. All results were within
proposed specification limits and no trends or significant changes were seen.

In addition, one batch was exposed to light as defined in the ICH Guideline on Photostability Testing of
New Drug Substances and Products. No change in tablet assay, water content, dissolution or degradation
was observed as a result of light exposure.

Based on available stability data, the proposed shelf-life of 30 months with no special storage conditions
for the finished product packaged in opaque PVC/PCTFE/PVC-aluminium blisters as stated in the SmPC
(section 6.3) are acceptable.

Adventitious agents

No excipients of human origin are used in the manufacture of the product. As regards animal origin, the
only component of animal origin is shellac which is used in the printing ink. Shellac is obtained from the
resinous secretion of insects. Insects are not a transmissible spongiform encephalopathy (TSE)-relevant
animal species, as defined in the "Note for guidance on minimising the risk of transmitting animal
spongiform encephalopathy agents via human and veterinary medicinal products (EMA/410/01 rev.3).
The magnesium stearate is of vegetable origin.

2.2.4. Discussion on chemical, pharmaceutical and biological aspects

Information on development, manufacture and control of the active substance and finished product has
been presented in a satisfactory manner. The results of tests carried out indicate consistency and
uniformity of important product quality characteristics, and these in turn lead to the conclusion that the
product should have a satisfactory and uniform performance in clinical use.

2.2.5. Conclusions on the chemical, pharmaceutical and biological aspects

The quality of this product is considered to be acceptable when used in accordance with the conditions
defined in the SmPC. Physicochemical and biological aspects relevant to the uniform clinical performance
of the product have been investigated and are controlled in a satisfactory way.

2.2.6. Recommendation(s) for future quality development

Not applicable.

Assessment report
EMA/178275/2019 Page 16/125
2.3. Non-clinical aspects

2.3.1. Pharmacology

Primary pharmacodynamics

In vitro studies demonstrated sotagliflozin is an inhibitor of human SGLT2 (IC50 = 1.8 nM) and SGLT1
(IC50 = 36.3 nM). Sotagliflozin also inhibits SGLT2 and SGLT1 from non-human species (mouse, rat, dog
and monkey). SGLT2 potency ranged from 0.5 - 1.4 nM, while SGLT1 potency ranged from 40.2 to 85.6
nM.

Through its activity at SGLT2 sotagliflozin enhanced glucose secretion in urine in mice, rats, dogs and
monkeys. In diabetic and non-diabetic mice and in obese-prone non-diabetic rats, sotagliflozin improved
glucose tolerance. In obese-prone non-diabetic rats, sotagliflozin lowered postprandial insulin. Through
inhibition of SGLT1 in the intestine, sotagliflozin reduced glucose absorption. Postprandial levels of GLP-1
and PYY were increased. A study in SGLT1 knockout mice showed that the increase in GLP-1 was not
dependent on SGLT1-inhibition. No increase in PYY was observed in the SGLT knockout mice. In
non-obese diabetic (NOD) mice with T1DM, sotagliflozin significantly improved glycemic control in
animals maintained on a low dose of insulin; fed blood glucose and A1C levels were comparable to mice
treated with high-dose insulin only. The incidence of hypoglycaemia was much lower in the
sotagliflozin-treated mice receiving low dose of insulin than in the mice receiving high-dose insulin only.

Secondary pharmacodynamics

No relevant off-target activity was observed in a general screen with 75 different receptors, enzymes and
ion channels. In an assay with other members of the SGLT family, relevant activity was observed with
hSGLT5 (IC50 = 54.1 nM) and hSGLT6 (IC50 = 27.3 nM).

Safety pharmacology

Sotagliflozin has an IC50 >23.5 µM in the human ether-a-go-go related gene (hERG) assay. A cardiac
telemetry study in Beagle dogs showed that sotagliflozin had no significant effect on QT or QTc interval
(corrected with the Fridericia method) at dose levels up to 500 mg/kg. Sotagliflozin at 500 mg/kg caused
significantly higher heart rates (51% higher than controls) in Beagle dogs. Increased heart rate (+18%)
was also noted for animals administered 100 mg/kg at two hr post-dose. Sotagliflozin at doses
≤100 mg/kg had no adverse effects on the CNS, respiratory, GI, and renal systems. As expected based
on the mode of action, sotagliflozin had pharmacologic effects on the renal system. These changes were
considered non-adverse at doses ≤100 mg/kg sotagliflozin.

2.3.2. Pharmacokinetics

Following single intravenous administration of sotagliflozin to mice, rats, dogs and monkeys, sotagliflozin
exhibited a generally similar volume of distribution, whereas clearance appeared to be slightly greater in
rodents than in the dog or monkey. Following administration of a single oral dose of sotagliflozin, good
oral bioavailability (≥50%) was observed in the mouse, rat, and dog. In the monkey, oral bioavailability
of sotagliflozin was lower (approximately 5%). In all species, exposure following oral administration of
sotagliflozin increased as a function of the dose administered. In multiple dose studies, sex-related
differences were observed with females exhibiting increased exposure when compared to the males.
There was no evidence of significant sotagliflozin accumulation on repeated oral administration.

Sotagliflozin exhibited high plasma protein binding (>90%) in all species tested. Quantitative whole-body
14
autoradiography (QWBA) studies in male rats administered C-sotagliflozin by oral gavage,
demonstrated that radioactivity was extensively distributed to tissues and organs. Tissues showing the
highest peak concentrations of radioactivity were the liver, kidney medulla, kidney cortex, kidney,

Assessment report
EMA/178275/2019 Page 17/125
adrenal gland, and pancreas. There was no significant accumulation of radioactivity in the central
nervous system (CNS), and radioactivity was not selectively associated with the melanin-containing
tissues or red blood cells. The placental transfer of sotagliflozin was assessed following administration of
14 14
C-sotagliflozin to timed pregnant rats. The results showed that C-sotagliflozin-related radioactivity
crossed the placenta at low levels.

Both in vitro and in vivo metabolism studies demonstrated that metabolic profiles are qualitatively similar
among species, except dog where minimal metabolism was observed. Sotagliflozin was extensively
metabolized in the in vivo human absorption, metabolism, and excretion study. Direct glucuronidation,
forming sotagliflozin-3-O-glucuronide, was the major metabolic pathway, accounting for 94% of total
radioactivity in plasma (0.25 to 48 hr AUC pool), and 33% of the dose in urine. The glucuronide of
sotagliflozin was present in all species except dog. All other metabolites detected in human plasma each
accounted for 2.5% or less of the total radioactivity.

2.3.3. Toxicology

Single dose toxicity

In a single dose toxicity study in rats, sotagliflozin was well tolerated at doses up to 1000 mg/kg.

Repeat dose toxicity

In the rat repeat dose toxicity studies, the expected pharmacologic effect of marked glucosuria was
observed. The glucosuria predisposed for infection and inflammation of the urogenital system. In the
26-week repeat-dose study in rats, glucosuria-related changes were observed at ≥75 mg/kg/day males
and 300 mg/kg/day females. Sotagliflozin-related targets included the prostate, trabecular bone of the
sternum, renal pelvis (females), and urinary bladder (females). The applicant refers to other marketed
SGLT2 inhibitors, where prostate inflammation was not observed. Based on an evaluation of available
nonclinical and clinical data as well as published literature, the applicant is of the view that this
observation is of limited relevance in humans. Sotagliflozin-related increase in trabecular bone (minimal
to moderate) was noted within the sternum of males (≥30 mg/kg/day) and females (≥75 mg/kg/day) in
the 26-week repeat-dose rat study and was characterized by a thickening of the trabecular bone adjacent
to the growth plate. An assessment of bone turnover markers in the 26-week rat study showed that there
was a decrease in bone turnover markers (osteocalcin, N-terminal propeptide of type 1 collagen, and
Osteoclast-derived tartrate-resistant acid phosphate) and calciotropic hormones (1,25 Vit. D and PTH).
The trend for a decrease in bone turnover markers at 30 mg/kg/day was generally marginal and not
considered biologically significant.

In the dogs, adverse effects were observed at the highest dose (500/300 mg/kg) in the 4-weeek and
13-week studies, In the 39 week study no adverse effects were observed at the highest dose 200 mg/kg.
All observations were considered related to the desired pharmacological effect. No changes in the bone
turnover markers and calciotropic hormones were observed in the 39-week dog study.

Genotoxicity

A standard battery of genetic toxicology studies (Salmonella-Escherichia coli/mammalian-microsome

reverse mutation assay (Ames assay); chromosomal aberration test in Chinese Hamster Ovary (CHO)
cells; and the in vivo rat bone marrow micronucleus assay) determined that sotagliflozin was not
mutagenic or clastogenic.

Carcinogenicity

The carcinogenic potential of sotagliflozin was assessed in transgenic (Tg.RasH2) mice and
Sprague-Dawley rats. In the 26-week mouse study, no sotagliflozin-related effects on survival or

Assessment report
EMA/178275/2019 Page 18/125
incidence of neoplasms were observed at doses up to 100 mg/kg/day, the highest dose evaluated. In the
104-week rat study, sotagliflozin-related proliferative changes (neoplasia and hyperplasia) were
observed in the thyroid gland, and sotagliflozin-related non-neoplastic microscopic findings were noted in
the kidney, urinary bladder, parathyroid, prostate, seminal vesicle, nonglandular and glandular stomach,
tongue, sternum, and femur. An increase in the incidence of thyroid follicular cell carcinoma was
observed for 75 mg/kg/day males. Sotagliflozin was not considered to cause carcinogenicity in rats at
≤30 mg/kg/day. In a mechanistic study it was shown that sotagliflozin significantly increased serum TSH.
Sotagliflozin increased the average frequency of Ki67-postive staining (a marker for cell proliferation) in
the thyroid. The applicant concludes that the thyroid carcinoma observed in the study were the result of
the TSH increase, known to increase follicular cell tumour incidence in rats by a mechanism that is not
relevant for humans.

Reproductive and developmental toxicity

Sotagliflozin did not impair fertility or alter early embryonic development in the rat at doses up to
300 mg/kg/day, the highest dose evaluated. Embryo-fetal development was evaluated in rats and
rabbits. In rats, the NOAEL for embryo-fetal development was 100 mg/kg/day, with marked maternal
and embryo-fetal toxicity at 350 mg/kg/day. In the rabbit, the embryo-fetal toxicity no-observed-effect
level (NOEL) was 200 mg/kg/day, the highest dose evaluated. In a rat pre- and post-natal development
study the NOAEL was 100 mg/kg/day, the highest dose administered. In a rat juvenile toxicity study, the
NOAEL was considered to be 75 mg/kg/day, the highest dose evaluated.

Toxicokinetics

Exposure multiples in relation to clinical exposure at the maximal recommended human dose is
summarized in the following table:

Study Type Species Sex NOEL/NOAEL AUC0-24h MOEa

(mg/kg/day) (ng*hr/mL)
M 30 15 077 8x
26-Week Repeat-dose Rat
F 75 84 145 43x
M 158 334 80x
39-Week Repeat-dose Dog 200
F 143 795 73x
M 87 400 44x
Mouse 100
F 207 000 105x
Carcinogenicity
M 14 800 8x
Rat 30
F 40 300 20x
Rat F 100 78 012 40x
Embryo-fetal toxicity
Rabbit F 200b 18 145 9x
M 34 500 18x
Juvenile toxicity Rat 75
F 60 100 31x
Abbreviations: F=female; M=male; MOE=margin of exposure; MRHD=maximum recommended human dose;
NOAEL= no-observed-adverse-effect level; NOEL=no-observed-effect level
a MOE = Margin of exposure. Compares exposure of sotagliflozin in animal plasma (AUC0-24h) at the NOEL/NOAEL
to the human AUC0-24h at the maximum recommended human dose (MRHD) of 400 mg dose (1969.3 ng*hr/mL;
Population PK study report LX4211-N101).
b NOEL value
Other toxicity

Process-related impurities were observed in sotagliflozin drug substance, however no individual

impurities that exceeded the ICH identification threshold of 0.10% were observed in clinical batches.

A mutagenic risk assessment was performed on actual and potential mutagenic impurities of sotagliflozin
according to the ICH M7 guideline.

No phototoxicity study has been performed since sotagliflozin does not absorb within the 300 to 800 nm
range of the electromagnetic spectrum. While weak absorption was observed at 290 nm, its molar

Assessment report
EMA/178275/2019 Page 19/125
extinction coefficient was below 1000 L mol-1cm-1 and is therefore not considered to be sufficiently
photoreactive to result in direct phototoxicity.

2.3.4. Ecotoxicity/environmental risk assessment

The phase I default PECSW for sotagliflozin was calculated to 2.0 x 10-3mg/L or 2.0μg/L, triggering a Phase
II assessment. The log KOW for sotagliflozin was reported as 2.98 but the study report provided
(#LP-802034-01-014) was of insufficient quality. The CHMP requested the study reports and an updated
ERA to be provided post-approval.

2.3.5. Discussion on non-clinical aspects

Pharmacology

The data on primary pharmacodynamics presented by the applicant suggest that sotagliflozin acts
through inhibition of SGLT1 and SLT2 to improve glycemic role. The inhibition of SGLT2 is the most
important mode of action. The contribution of SGLT1 inhibition was shown in the animal studies but it is
difficult to extrapolate this to the human situation. The 20-fold lower activity at SGLT1 results in that
inhibition of SGLT1 is likely to take place only locally in the intestine. In mice and rats, the activity at
SGLT1 is even lower (84-fold in mouse, 202-foldl in rat), suggesting a lesser role for SGLT-1 in the rodent
models. However, there are clear differences in the metabolism of sotagliflozin between species that
makes the comparison difficult. An estimate of the relative role of SGLT2 vs SGLT1 inhibition must be
made on clinical PD data.

No relevant off-target activity was observed in a general screen with 75 different receptors, enzymes and
ion channels. In an assay with other members of the SGLT family, relevant activity was observed with
hSGLT5 (IC50 = 54.1 nM) and hSGLT6 (IC50 = 27.3 nM). SGLT5 and SGLT6 act primarily in the kidney and
CNS, respectively. The activity at these targets is similar to the activity at SGLT1. Since systemic
inhibition of SGLT1 has not been observed, it unlikely that any inhibition of SGLT5 or SGLT6 will occur at
clinically relevant exposures.

Although a few compound-related findings were noted in these safety pharmacology assessments, it is
agreed that the dose levels associated with the findings were well in excess of those required to produce
the desired pharmacologic effects.

Pharmacokinetics

The major human metabolite, sotagliflozin-3-O-glucuronide constitutes the dominant form in human
plasma (94%). This metabolite is formed in the toxicity species, but to a much smaller extent. This is
important to consider in the safety evaluation.

Toxicology

Repeat-dose toxicity studies were performed in rats and dogs. Carcinogenicity evaluation was performed
in mice and rats. Embryofetal development studies were performed in rats and rabbits. All species are
pharmacologically relevant. While no data on the activity of sotagliflozin on rabbit SGLT1 and SGLT2 were
provided, maternal effects related to the expected pharmacology were observed. The main human
metabolite sotagliflozin-3-O-glucuronide was formed in rats and mice at much lower levels than in
humans and was not found in dog. Whereas this metabolite is not qualified in these animal studies it is not
considered a safety concern, based on experience with other glucuronides. The choice of animal species
for toxicity evaluation is endorsed.

Repeat dose toxicity

Assessment report
EMA/178275/2019 Page 20/125
In the rat repeat dose toxicity studies, the expected pharmacologic effect of marked glucosuria was
observed. The glucosuria predisposed for infection and inflammation of the urogenital system, including
prostate, renal pelvis and urinary bladder. The applicant refers to other marketed SGLT2 inhibitors, where
prostate inflammation was not observed. Based on an evaluation of available nonclinical and clinical data
as well as published literature, the applicant is of the view that this observation is of limited relevance in
humans. This conclusion is endorsed.

Sotagliflozin-related increase in trabecular bone (minimal to moderate) was noted within the sternum of
males (≥30 mg/kg/day) and females (≥75 mg/kg/day) in the 26-week repeat-dose rat study. An
assessment of bone turnover markers in the 26-week rat study showed that there was a decrease in bone
turnover markers. Bone mineral density and bone resorption biomarkers were monitored in clinical
studies. There were small changes observed in bone density versus placebo in a 52 week substudy, while
the bone density values were still within the normal range. In view of these minor effects, it is agreed that
the rat findings are of little clinical importance.

In all repeat-dose toxicity studies, NOAEL values did not consider the pharmacologic effect of glucosuria
as adverse. While it is generally not acceptable to disregard an adverse effect that is related to primary
pharmacology, and glucosuria would normally be considered an adverse effect, in this case such NOAEL
determination would not be meaningful. Glucosuria is the desired effect in the clinical situation. The
NOAEL values proposed by the applicant are endorsed.

Carcinogenicity

In the rat carcinogenicity state, an increase in the incidence of thyroid follicular cell carcinoma was
observed for 75 mg/kg/day males. In a mechanistic study it was shown that sotagliflozin significantly
increased serum TSH. It is agreed that these data support a mechanism where sotagliflozin causes an
increase in thyroid hormone concentrations which in rat is associated with a carcinogenic risk. Such effect
is considered not relevant in humans and it is therefore concluded that the rat findings do not suggest a
carcinogenic potential in humans.

Reproductive and developmental toxicity

A standard package on reproductive and developmental toxicity was performed in rat and rabbit. It is
agreed that these data do not suggest a risk for reproductive or developmental toxicity for sotagliflozin.
Reversible renal changes (kidney weight; tubular dilatation) were observed in the rat juvenile toxicity
study. Because glomerular development/nephrogenesis in humans is completed prenatally and the
timing of nephrogenesis in the rat occurs postnatally, the findings in the juvenile rat toxicity study are
relevant for exposure during the 2nd and 3rd trimester in humans. A recommendation against use during
this period is included in the SmPC section 4.6

Metabolites

The main human metabolite, sotagliflozin-3-O-glucuronide,is formed in small levels in the toxicity
species. The exposure levels in the animal studies have not been quantified, but they are unlikely to
provide a formal qualification. However, glucuronides other than acyl glucuronides are considered of little
safety concern and no further qualification studies are warranted. This was also the conclusion in a
previous CHMP Scientific Advice.

RMP

In the view of the CHMP, there are no nonclinical findings, which have not already been covered by clinical
safety data that should be raised as important potential risks in the RMP.

Environmental risk assessment

The CHMP requested the missing ERA files (phase II and log Kow study) to be provided post-approval.

Assessment report
EMA/178275/2019 Page 21/125
2.3.6. Conclusion on non-clinical aspects

This application was considered approvable from a non-clinical point of view. Remaining information
needed for finalising the Environmental Risk Assessment is requested post-approval.

2.4. Clinical aspects

2.4.1. Introduction

GCP

The Clinical trials were performed in accordance with GCP as claimed by the applicant.

The applicant has provided a statement to the effect that clinical trials conducted outside the Community
were carried out in accordance with the ethical standards of Directive 2001/20/EC.

• Tabular overview of clinical studies

Table 2 Overview of studies included in the clinical pharmacology package of

sotagliflozin
Description Phase Subject n Dose Reference

SAD 1 HV 96 5, 15, 25, 50, 100, 200, 300, 500 mg LX4211-1-101

50, 100, 200, 300 od - 1 week
MAD 32

SAD - high doses 1 HV 24 800, 1200, 2000 mg LX4211-1-106

MAD - high doses 1 HV 25 400, 800 mg od - 10 days LX4211-1-110

QT 1 HV 58 800, 2000 mg LX4211-1-109

Frel tablet vs solution 1 T2DM 15 300 mg LX4211-1-102

6x50mg, 2x150 mg, 30 ml 10 mg/ml

BE 2x200 mg vs 400 mg 1 HV 30 400 mg LX4211-1-117

BE commercial vs 1 HV 76 200 mg BEQ15271

development tablet

Food effect 1 HV 14 400 mg PKM15047

Mass balance 1 HV 8 400 mg 14C-sotagliflozin (100 µCi) LX4211-1-108

Renal impairment 1 T2DM 16 400 mg od - 1 week LX4211-1-107

(pk)

Renal impairment 1 RI 26 400 mg LX4211-1-121

Hepatic impairment 1 HI 32 400 mg LX4211-1-116

DDI - metformin 1 HV 30 400 mg LX4211-1-103

DDI - COC 1 HV 30 400 mg - 1 week LX4211-1-120

DDI - rifampicin 1 HV 16 400 mg INT14936

DDI - mefenamic acid 1 HV 16 400 mg od for1 week INT14937

DDI - digoxin 1 HV 24 400 mg od for 12 days LX4211-1-114

DDI- rosuvastatin 1 HV 24 400 mg od for 13 days LX4211-1-115

DDI - metoprolol 1 INT14972

midazolam

Assessment report
EMA/178275/2019 Page 22/125
Table 3 Overview of modelling and simulation reports included in the submission of
sotagliflozin
Description Reference

PopPK - healthy volunteers, CKD pats POH0532

PopPK and HBA1c exposure-response - T1D2 pats LX4211-N101
Pop exposure-response - T1D2 pats LX4211-N103

Assessment report
EMA/178275/2019 Page 23/125
Table 4 Phase 2 and 3 studies included in the clinical development program
Study Study Short Dosing Schema/ Patient Number of Status
Number Title/Design Study Duration Population Patients
CTD Randomized
Location to Treatment

Phase 2 Studies
LX4211.202 Phase 2, Multiple oral doses of T2DM with 299 total Initiated:
Placebo-controlled, sotagliflozin inadequate 15-Jun-2011
(Study 202) 59 in 75 mg qd;
Dose-Ranging (75 mg qd, 200 mg glycemic control
60 in 200 mg qd; Completed:
Study in qd, 400 mg qd; on metformin
60 in 400 mg qd; 15-May-201
Combination with 200 mg bid) or monotherapy
60 in 200 mg bid; 2
Metformin in T2DM placebo administered
60 in placebo qd.
for 12 weeks

LX4211.203 Phase 2, Sotagliflozin 400 mg Inadequately 36 total Initiated:

Placebo-controlled, or placebo controlled T1DM 08-Feb-2013
(Study 203) 19 in 400 mg qd;
Proof-of-Concept administered qd for
17 in placebo qd. Completed:
Study in T1DM 29 days
13-Jan-2014

LX4211.204 Phase 2, 400 mg sotagliflozin 18 to 30 87 total Initiated:

Placebo-controlled or placebo year-old adult 20-April-201
(Study 204) 43 in 400 mg qd;
Study in Young administered qd for patients with 5
44 in placebo qd.
Adult Patients with 12 weeks T1DM with A1C
Completed:
T1DM and Elevated ≥9%
23-Sep-2016
A1C

LX4211.206 Phase 2, Placebo- 75 mg, 200 mg, or T1DM and 141 total Initiated:
controlled, 400 mg administered inadequate 10-Jul-2015
(Study 206) 35 in 75 mg qd;
Dose-Ranging qd for 12 weeks glycemic control
35 in 200 mg qd; Completed:
Study in T1DM with insulin
35 in 400 mg qd; 26-Aug-2016
therapy
36 placebo qd.

Phase 3 Studies
LX4211.309a Phase 3, 200 mg or 400 mg T1DM 793 total Initiated:
Placebo-controlled sotagliflozin or North America 23-Mar-2015
(Study 309) 263 in 200 mg qd;
Study in T1DM placebo qd for 52
262 in 400 mg qd; Completed:
with Optimized weeks (24-week Core
268 in placebo qd 17-Feb-2017
Insulin Therapy Treatment; 28-week
Long Term Extension)

Assessment report
EMA/178275/2019 Page 24/125
Study Study Short Dosing Schema/ Patient Number of Status
Number Title/Design Study Duration Population Patients
CTD Randomized
Location to Treatment

LX4211.310a Phase 3, 200 mg or 400 mg T1DM 782 total Initiated:

Placebo-controlled sotagliflozin or Europe & Israel 21-May-201
(Study 310) 261 in 200 mg qd;
Study in T1DM placebo qd for 52 5
263 in 400 mg qd;
with Optimized weeks (24-week Core
258 in placebo qd. Completed:
Insulin Therapy Treatment; 28-week
23-Jun-2017
Long Term Extension)

LX4211.312b Phase 3, 400 mg sotagliflozin T1DM 1405 total Initiated:

Placebo-controlled or placebo qd for Worldwide 18-Sep-2015
(Study 312) 700 in 400 mg qd;
Study to Evaluate 24 weeks
705 in placebo qd Completed:
the Net Clinical
18-Apr-2017
Benefitc of
Sotagliflozin as
Adjunct to Insulin
Therapy in T1DM
a
For Studies 309 and 310, insulin therapy was optimized prior to randomization, as detailed in Section
1.3.4
b
For Study 312, insulin was not optimized prior to randomization.
c
Net benefit was defined as the proportion of patients with A1C <7.0% and no episode of SH and no
episode of DKA from randomization to Week 24.

2.4.2. Pharmacokinetics

Bioanalysis

A HPLC-MS/MS method for simultaneous determination of sotagliflozin and M19

(sotagliflozin-3-O-glucuronide) in plasma has been developed, pre- and within study validated.

A HPLC-MS/MS bioanalytical assay for determination of sotagliflozin and M19 in the urine has been
developed and pre-study qualified.

LC-MS/MS methods for determination of rifampicin, mefenamic acid, rosuvastatin, digoxin, metoprolol,
midazolam, ethinylestradiol and norelgestromin were also developed and validated for the intended
purposes.

Absorption
Sotagliflozin is characterized as a BCSII compound. At least 57% of the dose entered the systemic
circulation as parent compound/or metabolites. Following both single and repeated dosing with
sotagliflozin tablets, the Cmax occurred at 1-4h but secondary peaks/shoulders were seen in the
concentration-time profiles.

The increase in systemic exposure was dose-proportional in the therapeutic dose range. The
accumulation index at steady state was ca 2.

Bioequivalence was demonstrated between the commercial 200-mg tablet and the 200-mg development
tablet used in the clinical program including phase 3.

The exposure of sotagliflozin was increased when co-administered with food (high fat/high caloric), the
Cmax was ca 2.5-fold and total exposure ca 1.5-fold higher compared to under fasted condition.

Assessment report
EMA/178275/2019 Page 25/125
Distribution
The fu (unbound fraction) was determined in vitro to 2.7% and to 2.3% for sotagliflozin and M19,
respectively, in the therapeutic concentration range.

Ex vivo protein binding of sotagliflozin in plasma from subjects diagnosed with renal and hepatic
impairment did not show any differences in fu compared to in healthy subjects.

Elimination
The terminal t1/2 of sotagliflozin was calculated to about 24-30h and to about 24h for M19.

Fifty-seven percent of a 14C-sotagliflozin dose was excreted in the urine and 37% in faeces. Less than 1%
of the dose was excreted unchanged compound in the urine and 23% of the dose was identified as
sotagliflozin in faeces.

Metabolism
Sotagliflozin was extensively metabolised by UGT1A9 in vitro and UGT1A1 and 2B7 were also involved but
to a lesser extent. UGT1A9 is characterized as a polymorphic enzyme and genotyping for UGT1A9*3 was
performed in four phase-1 studies. One subject bearing the UGT1A9*3 allele was identified and exhibited
increased exposure of sotagliflozin.

In vitro phenotyping showed that sotagliflozin was also metabolised by CYP3A4.

A total of 22 metabolites were characterised based on pooled plasma, urine and faeces samples. Twelve
in the urine and 11 in faeces. Many of the metabolites were identified as glucuronides.

One major metabolite M19, a direct glucuronide (sotagliflozin-3-O-glucuronide) was identified,

representing 94% of the total plasma radioactivity. M19 was characterized as a very poor inhibitor with
IC50s of >10000 nM for both SGLT2 and SGLT1 transporters.

No active metabolites were identified.

Time-dependency
No signals of time-dependent PK of sotagliflozin have been seen following repeated dosing.

PK in the target population

The individual model predicted exposure of sotagliflozin at steady state following 200 mg od was Cave=40
ng/ml with Cmax=64 ng/ml and Cmin=29 ng/ml. The estimated time to peak concentration varied
considerably with a typical value of 2h post dose. The total exposure was AUCτ=968 ng/ml.h and the
terminal t1/2 ca 28h. The inter-individual variability was ca 47% (CV) and the intra-individual 9-32% and
the intra-individual 9-32%.

Special population
Renal impairment
The total exposure of sotagliflozin, and its variability, increased with decreasing renal function but levelled
off between moderate and severe RI. The exposure was about 1.7-, 2.7-fold, 0.8- and 1.5-fold higher in
mild, moderate and RI and ESRD, respectively, compared to healthy subjects. M19 represented >95% of
the total systemic exposure independently of degree of renal function.

No difference in the exposure of sotagliflozin was, however, seen between T2DM subjects diagnosed with
moderate (eGFR ≥45 ml/min/1.73m2) and severe (eGFR <45 ml/min/1.73m2) RI following 400 mg od for
one week. The exposure on Day 7 i.e. at steady state was ca 2.5-fold higher than after a single dose.

Hepatic impairment
The total exposure of sotagliflozin was comparable (slightly lower 0.7-fold) in subjects diagnosed with
mild HI with compared healthy subjects in a dedicated HI study. About 3- and 5.6-fold higher exposure of
sotagliflozin was seen in subjects with moderate and severe HI, respectively, compared to healthy

Assessment report
EMA/178275/2019 Page 26/125
subjects following single oral dose of 400 mg. Comparable exposure of M19 was seen in mild and severe
HI compared to healthy subjects, respectively.

Co-variates
Body weight was identified as a significant covariate for CL/F. The model predicted CL/F was about
3.5-fold higher in a heavy subject weighing 166 kg compared to a small subject weighing 46 kg. Any
increase in exposure of sotagliflozin with age was probably not due to increased age but rather to a lower
eGFR with increasing age. Gender and race were not identified as clinical relevant co-variates.

Interactions
The PK interaction potential of sotagliflozin has been evaluated in a number of in vitro and seven in vivo
studies. Sotagliflozin was extensively metabolised to M19 (sotagliflozin-3-O-glucuronide) which was
investigated in the DDI studies as well. The results are tabulated/summarized below.

In vitro
Signals were shown for clinical relevant
- inhibition of CYP2D6 and CYP3A4
- induction in vivo of CYP1A2, CYP2B6 and CYP3A4
- inhibition of Pgp, BCRP, OATP1B1 and OATP1B3

In vivo
Victim
Co-treatment with
− rifampicin (strong inducer) - SmPC 4.5: frequent monitoring
− mefenamic acid (UGT inhibitor) - SmPC 4.5: no special dose recommendations
− metformin or combined oral contraceptive - no effect on the PK

Assessment report
EMA/178275/2019 Page 27/125
Perpertrator
Co-treatment with sotaglflozin and
− digoxin (Pgp substrate) - SmPC 4.5: careful monitoring
− midazolam (CYP3A4 substrate) - SmPC 4.5: unclear mechanism, more data needed
− rosuvastatin (OATP, BCRP, OAT3 substrate) - SmPC 4.5: unclear
− metformin (OCT2 substrate), ethinyl estradiol (CYP3A4 substrate) and norgestimate (CYP3A4, UGT
substrate) or metoprolol (CYP2D6 substrate) - no effect on their PK

2.4.3. Pharmacodynamics

Mechanism of action

Sotagliflozin is a dual SGLT1 and SGLT2 inhibitor. Based on anatomical location of SGLT, the half maximal
inhibitory concentration and protein binding of sotagliflozin, the SGLT1 inhibition is limited to local in the
GI tract, and the SGLT2 inhibition is systemic.

In normoglycaemic individuals, kidneys filter up to 180 g of glucose daily, 99% of which is reabsorbed into
the circulation primarily by membrane-associated proteins of the SGLT family located in the proximal
tubules of the nephron. This process results in little to no loss of glucose into the urine. When blood
glucose (BG) concentrations exceed the renal threshold (approximately 150 to 200 mg/dL), which
commonly occurs in diabetes, the maximum transport capability of SGLT2 is exceeded, resulting in
incomplete glucose reabsorption and glucose loss into the urine. SGLT2 inhibition in the kidney has
several consequences: 1) glucose is cleared from the circulation independently from insulin; 2) glucose
clearance decreases with reduced levels of BG, which limits the risk of severe hypoglycaemia; 3) urinary
glucose excretion (UGE) lowers blood pressure (BP) through mild diuresis; and 4) UGE leads to weight
loss through caloric loss (2).

The primary transporter for absorption of glucose and galactose in the intestine is SGLT1. Inhibition of
SGLT1 delays and reduces glucose absorption in the proximal intestine, resulting in blunting and delay of
postprandial glucose (PPG) excursions, and decreased glycaemic variability. In addition, more glucose
delivery distally triggers increased L-cell release of glucagon-like peptide-1 (GLP-1) and peptide YY (PYY),

Assessment report
EMA/178275/2019 Page 28/125
which is a natural peptide hormone response associated with enhanced glycaemic and appetite control.
Thus inhibition of SGLT1 in the intestine may improve glucose control in several ways, namely: 1)
reducing intestinal glucose absorption leading to reduced PPG; and 2) stimulation of GI peptides, such as
GLP-1 and PYY, proteins that assist in glycaemic and appetite control, and are associated with reductions
in BP and body weight (BW).

With sotagliflozin, the desirable effects of SGLT2 inhibition are complemented with SGLT1 inhibition in the
GI tract to produce reductions in PPG while triggering lower UGE than the more selective SGLT2 inhibitors.
The dual SGLT1 and SGLT2 inhibitory activity of sotagliflozin could potentially maintain efficacy in patients
with renal impairment as well as reduce the side effects associated with high UGE, such as genitourinary
tract infections.

Primary pharmacology

The PD of sotagliflozin has been studied in healthy subjects (LX4211.104, LX4211.111) and subjects with
T2DM with/without concomitant sitagliptin (LX4211.105), at a single dose level of 400 mg (LX4211.105
and LX4211.111) or multiple dose regimens of 150 mg, 300 mg, or 400 mg qd (LX4211.201,
LX4211.104). Main PD markers in these studies were PPG and UGE. In the glucose tracer study
LX4211.111, the primary endpoint was RaO and was compared to canagliflozin. Among the further PD
markers FPG, GLP-1 (total and active), and PYY were common across studies.

A Phase 2 study in the target population (LX4211.203) investigated a wide range of PD endpoints in T1DM
patients with primary focus on the reduction in the bolus insulin doses under a dose regimen of 400 mg
sotagliflozin qd or placebo over 4 weeks. PD data has further been provided from the dose-ranging study
in T1DM (LX4211.206).

In addition limited PD data was provided from a study in patients with T2DM and renal impairment
(LX4211.107).

Urinary glucose excretion

Urinary glucose excretion was increased in all populations tested, consistent with SGLT2 inhibition. After
a single dose of 400 mg sotagliflozin in healthy volunteers, the amount of glucose excreted in urine over
24 hours after dosing ranged between 40 and 45 g glucose in LX4211.111, which was lower than what
was seen after 300 mg of canagliflozin, a more selective SGLT2 inhibitor. After a single dose of 400 mg
sotagliflozin given to patients with T2DM in LX4211.105, the 24-hour UGE was approximately 92 g, which
constituted a change from Baseline of approximately 68 g.

Urinary glucose excretion was also assessed following multiple dosing of the liquid formulation in T2DM
patients treated over 4 weeks at 2 dose levels vs. placebo (LX4211.201). The Baseline Day -1 values
ranged from 13.8 to 16.6 g for the 3 dose groups. LS means of changes from baseline were around 46 and
66 g/day for the sotagliflozin 150 mg and 300 mg dose groups, respectively, on Day 1. This effect slightly
decreased until Day 28. No change was observed in the placebo group. Patients with T2DM and moderate
renal impairment have shown a slightly smaller change from baseline in 24-hour UGE of around 39 g/day
on Day 7 of a multiple dosing regimen in study LX4211.107.

The change from baseline in 24-hour UGE in the target population of T1DM patients was investigated in
the dose-ranging study LX4211.206. This study showed a dose-dependent mean increase in UGE from
baseline at Week 12 in all of the sotagliflozin groups compared with placebo (Table 5).

Assessment report
EMA/178275/2019 Page 29/125
Table 5 ANCOVA Analysis of Change from Baseline in Urinary Glucose Excretion at Week 12 –
mITT Population, study LX4211.206

Placebo Sotagliflozin Sotagliflozin Sotagliflozin

(N = 36) 75 mg 200 mg 400 mg
(N = 35) (N = 35) (N = 35)
Urinary Glucose Excretion (g/day)

Baseline

N 28 30 26 29

Mean 6.9 15.8 6.6 10.0

Change from Baseline at Week 12

N 23 24 23 24

LS Mean (SE) 0.26 (10.54) 42.0 (10.52) 58.0 (10.51) 70.7 (10.29)

95% CLs for Change from (-20.7, 21.2) (21.1, 62.9) (37.1,78.9) (50.3, 91.1)
Baseline

p-value 0.98 <0.001 <0.001 <0.001

Summary of treatment comparison

LS Mean Difference (SE) from -- 41.8 (15.0) 57.7 (14.8) 70.4 (14.7)
Placebo

95% CLs for Difference -- (12.0, 71.5) (28.3, 87.2) (41.3, 99.6)

p-value -- 0.006 <0.001 <0.001

Post Prandial Glucose

The 2-hour PPG for both 150 mg and 300 mg sotagliflozin groups in T2DM patients in study LX4211.201
showed a lower level compared to the placebo group as early as Day 1 (p <0.05). At Day 28, the mean
change from Baseline in 2-hour PPG was -50.6 and -57.6 mg/dL in the 150 and 300 mg dose groups,
respectively, and were both statistically significant versus placebo (p <0.05).

LX4211.201 also included an evaluation of PPG changes from Baseline by analysing the AUC over 4 hours
after OGTT. In both the 150 mg and 300 mg dose groups, plasma glucose AUC0-4h corrected for baseline
and baseline-uncorrected AUC0-4h were statistically significantly lower than in the placebo group (p
<0.001). Reduction of plasma glucose AUC0-4h was observed at Day 27 of dosing to be approximately
35% lower compared to baseline (Day -2).

In study LX4211.107 in patients with T2DM and moderate to severe renal impairment, the equivalent
parameter of PPG, AUC(predose-4), confirmed a significant reduction from baseline with -176 mg*hr/dL on
Day 7 of treatment with 400 mg sotagliflozin, which constituted a change from baseline of around -20%.

In the target population of T1DM patients, the lowering of the 2-hour PPG showed a linear trend for dose
dependency in study LX4211.206 with the 400 mg qd dose level being statistically significantly different
from placebo with a change from baseline of -49.5 mg/dL at Week 12. The dose levels of 75 mg and 200
mg qd reached changes from baseline -20.3 and -27.4 mg/dL, respectively (not significant). Of note,
these values were observed while the patients (CSII group) had dose-dependently reduced their total

Assessment report
EMA/178275/2019 Page 30/125
daily bolus insulin doses on average by -2.85, -1.32 and -6.97 IU/day for the 75, 200 and 400 mg qd dose
levels, respectively.

In study LX4211.203, PPG was assessed by analysing the 3-hour post-breakfast glucose AUC (AUC0-3h) in
T1DM patients following a MMTT, performed on Days 1 and 29. In the sotagliflozin group (400 mg qd) the
mean change was -250 ± 187 mg*hr/dL which constituted a nearly 30% reduction compared to baseline,
compared with -200 ± 318 mg*hr/dL for the placebo group which was around 20% lower than baseline.
The difference between treatment groups was statistically significant (p = 0.009). This occurred despite
the reduction from Baseline in total daily bolus of insulin of 32.0 ± 18.1% in the sotagliflozin group, which
was a significantly greater decrease than for the placebo group (6.36 ± 27.7%).

The triple glucose tracer study, LX4211.111 (healthy volunteers), enabled a more differentiated view on
PPG by analysing oral and total glucose appearance after breakfast and lunch, as well as in comparison to
canagliflozin, a more selective SGLT2 inhibitor. Both drugs given as single doses prior to breakfast
reduced oral glucose appearance in the first 2 hours after a standard breakfast as compared to placebo,
but the sotagliflozin effect remained visible also at lunch whereas under canagliflozin it was not present
anymore at this time (Figure 2). On the other hand the oral glucose absorption after sotagliflozin was
slightly higher than under placebo between 2 and 5 hours after start of the test meals. Overall, this
resulted in a reduction in PPG excursions as also detected by CGM.

Figure 2 - Mean rate of oral glucose appearance from plasma [6,6-2H2] glucose enrichments
vs. time from test meal by treatment (lunch meal test group: completers population)

Abbreviation: SEM: standard error of the mean

The differences of the PPG AUCs were evaluated in a multiple dosing study (104) over 12 days in healthy
volunteers comparing on days 8 through 12 different dosing schedules relative to meal. In this study the
PPG AUCs were compared over the 3 main meals of the day. The outcome favoured a dosing schedule
immediately prior to breakfast, which was selected as the dosing schedule for the further Phase 2 and
Phase 3 studies.

Fasting plasma glucose

Study LX4211.104 in healthy volunteers showed a decrease from Baseline in FPG with all sotagliflozin
dose schedules, ranging from 5.17 mg/dL to 8.19 mg/dL, (p <0.001 for all dose groups).

Assessment report
EMA/178275/2019 Page 31/125
After 4 weeks of treatment in the T1DM study LX4211.203, sotagliflozin was not significantly different
from placebo in lowering fasting plasma glucose, but interpretation was limited by high variability of the
data. In the dose ranging study in T1DM, LX4211.206, when the LS mean for each group was compared
with the placebo group, the change in FPG from Baseline at Week 12 was not statistically significant at all
sotagliflozin dose levels and no dose-dependent effect was observed. This study was, however, difficult to
interpret due to group differences in the baseline FPG levels.

In the T2DM study LX4211.201, the decrease in mean FPG was observed starting on Day 7 in both
sotagliflozin dose groups. The lowest mean FPG values were 123.0 mg/dL for the sotagliflozin 150 mg
dose group and 120.3 mg/dL for the sotagliflozin 300 mg dose group on Day 29, after baseline values in
the range from 175.3 to 192.4 mg/dL. The mean FPG levels for the placebo group were 179.3 to 201.3
mg/dL and remained relatively stable throughout the study. Both sotagliflozin dose groups had
statistically significant lower FPG levels compared to that in the placebo group as early as Day 7
post-treatment through discharge on Day 29 (p <0.001). A trend towards a greater reduction in the
sotagliflozin 300 mg dose group was observed; however, the difference between the 2 sotagliflozin dose
groups was not statistically significant for any day of evaluation.

Incretins and gut hormones (glucagon-like peptide-1, peptide YY)

In study LX4211.104 in healthy volunteers, increases in GLP-1 (active) AUC0-last were seen in all
sotagliflozin dose schedules. In addition, all sotagliflozin dose schedules lead to increases from Baseline
in LS mean AUC0-last values of PYY.

The Phase 2 study LX4211.203 in T1DM patients assessed GLP-1 and PYY on Day 1 and Day 29. The
changes in active GLP-1 from Day 1 to Day 29 at 1-hour post-meal were 4.6 ± 4.0 pmol/L for the
sotagliflozin group compared with 1.8 ± 2.0 pmol/L for the placebo group. The changes in total GLP-1
from Day 1 to Day 29 at 1-hour post-meal were 7.5 ± 4.9 pmol/L for the sotagliflozin group compared
with 3.2 ± 2.3 pmol/L for the placebo group. Also, there were significant differences between the 2
treatment groups for the change in PYY. The changes from Day 1 to Day 29 at 1-hour post-meal were 9.8
± 4.8 pmol/L for the sotagliflozin group compared with 4.5 ± 5.9 pmol/L for the placebo group (LS mean
difference between groups 8.2 pmol/L, 95% CI: 4.5, 11.8; p <0.001).

Secondary pharmacology

QT-study (LX4211.109)

Study LX4211.109 evaluated the ECG effects of a single dose of sotagliflozin 800 mg and 2000 mg
compared with placebo and open-label moxifloxacin in 63 healthy subjects. The placebo-subtracted
change of QTcI from Baseline for both doses of sotagliflozin showed only slight elevation, primarily in the
first 6 hours post-dose, and all 1-sided 95% UCB were less than 10 msec compared to 10.76 msec for
moxifloxacin at 2 hours post-dose. The maximal mean value of ΔΔQTcI for the sotagliflozin 2000 mg dose
group was at 3-hours post-dose, where the ΔΔQTcI was 2.74 msec and the UCB was 4.50 msec. The
sotagliflozin 800 mg dose group produced a maximal change at 2-hours post-dose of 1.93 msec and an
UCB of 3.69 msec.

The steady state Cmax and AUC0-24h in T1DM patients at the maximum recommended human dose of 400
mg is 130 ng/mL and 1969 ng*hr/mL, respectively. In study LX4211.109, the Cmax and AUC0-last (last
measurable concentration was the last sampling time point: 60 hr) for the 800 mg and 2000 mg dose
were 139 ng/mL and 2257 ng*hr/mL; and 290 ng/mL and 4440 ng*hr/mL, respectively, indicating that
adequate coverage was achieved.

Assessment report
EMA/178275/2019 Page 32/125
Relationship between plasma concentration and effect

Pharmacokinetic/pharmacodynamic relationships related to efficacy

Population exposure-response (E-R) analyses were conducted to characterize the relationship between
sotagliflozin exposure and A1C, BW and 2-hour PPG.

Exposure-response analysis on Haemoglobin A1C

The exposure-A1C analysis aimed to characterize the relationship between the exposure of sotagliflozin
and A1C as a function of time. The covariates evaluated were: age, race/ethnicity, sex, BW, body mass
index, eGFR, renal function category, baseline and screening A1C and mean total daily insulin.

The exposure-response model on A1C was best described by a sigmoid inhibitory maximum
pharmacologic effect (Imax) model of time with a higher Imax (i.e. typical value of the effect on A1C
response) for subjects with screening A1C values (>10%) and with higher Imax values as weekly AUC of
sotagliflozin increases. Furthermore, the rate of increase in Imax related to drug exposure was found to be
higher for mean total daily insulin doses less than 0.73 IU/day/kg and lower for mean total daily insulin
dose greater than 0.73 IU/day/kg. The disease progression component of the model offsets the inhibition
of A1C, as compared to the predicted A1C, for placebo and drug effect.

The response on A1C was shown to be best predicted by using weekly AUC (sum of 7AUC0-tau) of
sotagliflozin as a marker for exposure. Model-predictions of the exposure-response relationships at week
24 showed that the effect was close to its maximum level at a weekly AUC of around 12000 ng*hr/mL. The
mean weekly AUC is simulated to be approximately 6999 and 14081 ng*hr/mL for 200 and 400 mg qd
doses of sotagliflozin, respectively. Furthermore, the model predicted that subjects treated with
sotagliflozin and total daily insulin doses <0.5 IU/kg had a decrease of 0.50% in A1C as compared to a
decrease of 0.59% for subjects with total daily insulin doses >1 IU/kg. Overall, the effect on A1C was
shown to be markedly influenced by a placebo effect, a treatment effect (i.e. sotagliflozin exposures),
baseline A1C and to a lesser extent by mean total daily insulin doses, as visualised in Figure 3. Of note,
eGFR was not found to be a significant covariate in the model.

Assessment report
EMA/178275/2019 Page 33/125
Figure 3 - LX4211-N101: Model-predicted Mean (90% Prediction Interval) of Change in A1C
From Baseline at Week 24, Stratified by Statistically Significant Covariate

Abbreviations: AUC, area under the plasma concentration-time curve; n, number of subjects in each group.

Exposure-response analysis on body weight and on 2-hour postprandial glucose

This exposure-response analysis aimed to characterize the relationship between the exposure of
sotagliflozin and BW or 2-hour PPG as a function of time. The covariates evaluated were: age, race,
ethnicity, sex, BW, body mass index, eGFR, renal function category and mean total daily insulin.

Exposure-response model on body weight

The exposure-response model on BW was described by 2 exponent of time functions: 1 for the placebo
effect and 1 for the treatment effect. The placebo effect was well described using a shift power function of
time with observed screening BW as the starting value and the exponent of the time function described as
a power function of age. The exponent describing the treatment effect was best described with an Emax
function of the weekly AUC. The decrease in BW was shown to begin plateauing at approximately 32
weeks postdose.

The exposure-response model, on average, resulted in an increase in BW of 0.24 kg for placebo by week
24 as compared to a decrease of 1.1 to 2.4 kg for active treatments with exposures as predicted for the
subjects of the Phase 3 studies. The probability of having an average decrease in BW of 2.2 to 2.4 kg was
predicted to be 7.3% for 200 mg and 43% for 400 mg of sotagliflozin. The model also predicted a
decrease in BW of 0.05 kg in placebo subjects ≥58 years as compared to an increase of 0.40 kg for
placebo subjects ≤32 years. In addition, the model predicted a higher decrease in BW in placebo subjects
with screening BW ≥99 kg (0.40 kg) as compared to placebo subjects with screening BW ≤68 kg (0.15
kg). As a result, subjects treated with sotagliflozin were predicted to have an 86% higher decrease in BW

Assessment report
EMA/178275/2019 Page 34/125
in the oldest (2.2 kg) as compared to the youngest (1.2 kg) and; a 133% higher decrease in BW for the
heaviest (2.3 kg) as compared to the lightest (1.0 kg). The effects of baseline body weight and age on the
change of body weight is visualised in Figure 4. Of note, eGFR was not found to be a significant covariate
in the model.

Figure 4 - LX4211-N103: Means and 90% Prediction Intervals of the Change in Body Weight
From Baseline After 24-weeks of Treatment, by Significant Covariate Groups and Dose Level

Abbreviations: AUC, area under the plasma concentration-time curve; n, number of patients.

Exposure-response model on postprandial glucose

The exposure-response model on the change in 2-hour PPG from baseline was described by a decreasing
exponential function of the model-predicted Cmin on the day of the 2-hour PPG measurement. This model
predicted the 2-hour PPG for placebo subjects to increase 3.6 mg/dL from baseline and predicted for
patients on active treatment the maximum decrease in 2-hour PPG (as Cmin approaches infinity) to be
44.7 mg/dL from baseline. Furthermore, 90% of the maximum decrease in 2-hour PPG from baseline was
predicted to occur at a Cmin value of approximately 20 ng/mL. Of note, eGFR was not found to be a
significant covariate in the model.

The model-predicted decrease in 2-hour PPG from baseline was 42 and 45 mg/dL by Week 24 at the
approximate median Cmin values of 26 and 46 ng/mL for 200 mg and 400 mg sotagliflozin, respectively.
The percentage of patients with predicted Cmin >23 ng/mL was approximately 60% for 200 and 88% for
400 mg sotagliflozin. Thus, the probability of a patient experiencing a greater than 40 mg/dL decrease in
2-hour PPG from baseline is 47% higher for 400 mg, as compared to a 200 mg sotagliflozin. At Cmin values

Assessment report
EMA/178275/2019 Page 35/125
of 20 ng/mL, the model-predicted the effect to be close to its maximum level (Figure 5). Median Cmin
values for 200 and 400 dose of sotagliflozin is of 42 and 45 ng/mL, respectively.

Figure 5 - LX4211-N103: Model-predicted changes from baseline in 2-hour PPG versus Cmin of
sotagliflozin overlaid with observed data from studies LX4211.204 and LX4211.206 (at week
12), LX4211.309 and LX4211.310 (at week 24)

Abbreviations: Cmin, drug concentration at 24 hours after a dose.

Note: The line represents the model predicted change from baseline

Overall, these results indicate a small difference between 200 and 400 mg dose of sotagliflozin on the
average model-predicted effect for A1C, BW and 2-hour PPG. Active subjects in the 400 mg dose group
had a higher probability of having a slightly larger decrease in A1C than subjects in the 200 mg dose
group.

Pharmacokinetic/pharmacodynamic relationships related to safety

Safety analyses were conducted to describe the probability of occurrence of positively adjudicated DKA
and severe hypoglycaemia as a function of sotagliflozin exposure. In these analyses, the influence of
different markers of sotagliflozin exposure at steady state (i.e. AUC0-tau, Cmin, and Cave) was evaluated.

Exploratory graphical analyses of the relationship between measures of steady-state sotagliflozin

exposure and the occurrence of DKA and severe hypoglycaemia showed that median AUC0-tau for patients
experiencing DKA is higher than in patients where DKA was not observed. Median AUC0-tau is similar in
patients with or without severe hypoglycaemia.

The effect of each sotagliflozin exposure on the probability of DKA and severe hypoglycaemia was
formally evaluated using linear logistic regression models. For DKA, AUC0-tau was found to be a significant
predictor, but the validity of the model-predicted probabilities could not be determined. For adjudicated
severe hypoglycaemia, no statistically significant exposure-response relationship was detected.

2.4.4. Discussion on clinical pharmacology

Pharmacokinetics

Following both single and repeated dosing with sotagliflozin tablets, the Cmax occurred at 1-4h but
secondary peaks/shoulders were seen in the concentration-time profiles, which might be an indication on
enterohepatic circulation.

Assessment report
EMA/178275/2019 Page 36/125
Sotagliflozin is characterized as a BCS II compound i.e. low solubility and good permeability. An oral
solution was used in the early phase-1 studies and co-administration with food resulted in a lower
exposure of sotagliflozin compared to in fasted state. This is expected when dissolution is not limiting but
delayed motility due to the concomitant food intake. The food effect was the opposite when the
development 200-mg tablet was administered together and without food, with a 1.5-fold higher exposure
in fed state. In vitro solubility data show that the local concentration of sotagliflozin in the intestine will be
well below concentrations inhibiting local transporter proteins.

All phase-1 studies have been performed in fasted state. In phase 3 sotagliflozin was administered before
the first meal of the day which is also the dose recommendation in the SmPC. Thus the clinical
pharmacology of sotagliflozin has been evaluated at lower concentrations (fasted condition) than in the
pivotal clinical studies. The phase-1 studies have been performed with a single dose of 400 mg and not
200 mg which is the recommended standard dose, 400 mg od should be used just in case of insufficient
response. The clinical pharmacology package is deemed satisfactory, considering the DDI profile for
sotagliflozin with no potential of dramatic interactions.

The elimination of sotagliflozin has not been completely elucidated with only ca 35% characterized, 33%
was excreted as M19 (direct glucuronide) and ca 1% excreted as parent compound in the urine. Thus 65%
of the elimination pathways are unknown. Thirty-seven percent of an oral 14C-sotagliflozin dose was
excreted in faeces with 23% as sotagliflozin. However, it is unknown if the excreted sotagliflozin is not
absorbed sotagliflozin, sotagliflozin excreted via the bile or sotagliflozin formed as a degradation product
from instable metabolites eg glucuronides. Biliary excretion seems as a potential elimination pathway. In
vitro stability data show that sotagliflozin-glucuronides are completely hydrolysed to sotagliflozin in the
presence of E.coli or bovine β-glucuronidase and M19 is characterized as a BCRP and MRP2 substrate, two
transporters known to be involved in the biliary excretion. An alternative way to understand the
elimination pathways is an in vivo DDI study with a known CYP3A4 inhibitor to characterize its importance
in vivo as CYP3A4 was identified in vitro as an enzyme involved in the metabolism of sotagliflozin. But
adding together small phase-1-metabolites, of which some was further conjugated, results in that
ca 20% of the dose was excreted via that way.

UGT1A9 is characterized as one of the main enzymes involved in the metabolism of sotagliflozin.
Genotyping for UGT1A9*3 allele was performed in four of the phase-1 studies, one subject bearing the
heterozygous UGT1A9*3 variant was identified and exhibited increased exposure. The in vivo DDI study
with co-administration of mefenamic acid (UGT inhibitor) resulted in a 1.5-fold higher exposure of
sotagliflozin when dosed together compared to when administered alone. This 1.5-fold increase is not
deemed clinical relevant resulting in any changes in the dose regimen. As sotagliflozin is an adjunct to
insulin and subjects are monitored at start of treatment, no change in the proposed dosing is considered
necessary for subjects bearing heterozygous UGT1A9*3 allele. Potential consequences (decreased
exposure) of other alleles with higher prevalence and associated with increased expression of UGT1A9 is
neither deemed clinical relevant with respect to changed dosing regimen as subjects are monitored at
start of treatment.

Adequate methods have been used in the development and evaluation of the popPK analysis of phase 1-3
data in T1DM patients and healthy volunteers. The structural model described the data sufficiently well.
The double peak behaviour seen in phase 1 data has not been captured, but is acceptable since it is not
foreseen that this simplification will impact the covariate analysis or the exposure-response analysis. Only
4% of the subjects in the dataset were classified with moderate renal impairment which could make the
influence of eGFR on CL/F for the lower range of the observed eGFR values less certain. Adequate
methods were used in the development and evaluation of the popPK analysis in healthy volunteers and
chronic kidney disease patients. The model described the data fairly well, although some bias was visible
for low concentrations, however, the results of the analysis are in agreement with the basic analysis of the
renal impairment studies and further refinement will not be pursued.

Assessment report
EMA/178275/2019 Page 37/125
The PK of sotagliflozin in T1DM patients were characterised by PopPK analysis. The individual predicted
exposure of sotagliflozin at steady state following 200 mg od was Cave=40 ng/ml. The estimated time to
Cmax varied considerably with a typical value of 2h post dose. The inter-individual variability was ca 47%
(CV) and the intra-individual to between 9-32%.

The total exposure of sotagliflozin, and its variability, increased with decreasing renal function and was
about 1.7-, 2.7-, 0.8- and 1.5-fold higher in mild, moderate and severe RI and ESRD, respectively, than
in subjects with normal renal function. The increase in exposure levelled off between moderate and
severe RI. The renal excretion of unchanged compound was <1% following an oral 14C-sotagliflozin dose
and the reason to the reported increase in systemic exposure in RI compared to in healthy subjects is
unclear. The interplay of renal and biliary excretion of sotagliflozin glucuronides and their intestinal
back-conversion to parent drug and enterohepatic circulation appears a reasonable explanation to the
unexpected increase in exposure with decreasing renal function for a compound excreted <1%
unchanged in the urine.

The PK interaction potential of sotagliflozin has been evaluated in a number of in vitro and seven in vivo
studies. Signals of inhibition of both CYP isozymes (CYP2D6 and 3A4) and transporter proteins (Pgp,
BCRP, OATP1B1 and OATP1B3) were shown in vitro for sotagliflozin or M19. M19 also showed in vitro
signals for induction of CYP1A2, CYP2B6 and CYP3A4. No inhibition of UGT isoenzymes was seen in vitro
at clinically relevant concentrations.

An increase in total exposure and Cmax of rosuvastatin of ca 1.2- and 1.4-fold, respectively, was seen
when co-administered with sotagliflozin. This is not deemed clinical relevant. However, the mechanism
behind the small increase in exposure is unclear as sotagliflozin and M19 (sotagliflozin glucuronide) are
characterized as BCRP inhibitors in vitro and M19 also as an inhibitor OATP1B3 and OAT3. Rosuvastatin is
a known OATP, BCRP and OAT3 substrate. It cannot be ruled out that sotagliflozin may interact with other
sensitive OATP- and/or BCRP-substrates resulting larger increases of exposure than seen for
rosuvastatin.

A single-dose DDI study with metformin, a known OCT2 substrate, was performed. Comparable systemic
exposure of metformin was seen when dosed together with sotagliflozin compared to when dosed alone.
Thus sotagliflozin/M19 are not an OCT2 inhibitor at clinically relevant concentrations. However,
metformin has dose-dependent PK and how representative the single-dose data are for the clinical
situation is not clear. However, metformin has a relatively short t1/2, ca 5h, and does not exhibit
accumulation. Thus, a multiple dose administration of metformin is not expected to impact the PK of
sotagliflozin in a significant different way compared to the single dose administration.

Sotagliflozin was a CYP3A4 inhibitor in vitro, with an IC50 of clinical relevance for a DDI in the GI tract.
M19 is identified as an inducer. The current in vivo DDI results show a decrease in exposure of midazolam
when co-administered with sotagliflozin of ca 0.8-fold compared to when dosed alone. Thus the net effect
of CYP3A4 inhibition + induction is 0.8-fold which leaves the maximum induction potential unknown.

Pharmacodynamics

The mechanism of action for SGLT2-inhibitors is well known as sotagliflozin is the fifth product in this
class. Sotagliflozin also exerts inhibition of SGLT1, predominantly expressed in the gut. This is predicted
to provide reductions in PPG through less glucose absorption after a meal. Furthermore, GI peptides are
stimulated which may further assist in glycaemic and appetite control. The applicant expect that the dual
effect on SGLT2 and SGLT1 will result in a lower UGE (due to lower absorption of glucose in the gut) which
in turn is expected to reduce side effects such as genitourinary infections and also will result in a
measurable effect also in patients with renal impairment.

SGLT2 inhibitors trigger multiple mechanisms that could predispose to DKA. When they are combined
with insulin, it is often necessary to decrease the insulin dose to avoid hypoglycaemia. The lower dose of

Assessment report
EMA/178275/2019 Page 38/125
insulin may be insufficient to suppress lipolysis and ketogenesis. Furthermore, SGLT2 is expressed in
pancreatic α-cells, and SGLT2 inhibitors promote glucagon secretion which stimulates lipolysis. Finally,
phlorizin, a nonselective inhibitor of SGLT family transporters, decreases urinary excretion of ketone
bodies. A decrease in the renal clearance of ketone bodies could also increase the plasma ketone body
levels.

Primary pharmacology of sotagliflozin was investigated in healthy subjects (studies 104 and 111),
patients with T2DM (studies 201 and 105) and patients with T1DM (study 203 and 206). One study
included patients with impaired renal function (107). The PD markers considered of most importance and
that were included as endpoints are UGE and PPG, considering sotagliflozin’s mechanism of action. In
addition the effect on FPG and incretins (GLP-1 (total and active) and PYY) were evaluated.

Studies 111 and 105 were single dose studies whereas the remaining studies investigated dosing over 2
to 4 weeks. In study 107, sotagliflozin was administered for 7 days.

Across the populations investigated, sotagliflozin resulted in an increase in UGE which appeared
dose-dependent (study 201, T2DM and study 203, T1DM). The magnitude of the effect differed between
populations as expected since the UGE is dependent on the blood glucose level, with the most prominent
effect observed in T2DM patients.

In all populations investigated (healthy volunteers, T2DM and T1DM) a consistent decrease of PPG was
observed with sotagliflozin treatment. As for UGE, the magnitude of the effect varied between populations
but the effect was larger with higher doses. The post-meal plasma glucose AUC0-4h/AUC0-3h decreased
with about 30% to 35% in both T1DM and T2DM patients with the higher dose.

When compared to canagliflozin, which also inhibit both SGLT2 and SGLT1 (study 111), the effect on PPG
appeared to be sustained for a longer time with sotagliflozin than with canagliflozin. This may indicate a
stronger SGLT1 inhibition with sotagliflozin. When compared to canagliflozin in healthy subjects, UGE was
lower with sotagliflozin than with canagliflozin. The applicant proposes that this is due to a stronger effect
on SGLT1 with sotagliflozin, which leads to less glucose being absorbed in the gut.

Concerns were raised that it could be more difficult to treat hypoglycaemia in patients under treatment
with a SGLT1 inhibitor. However, as seen in Figure 2, glucose appearance is not affected by the SGLT1
inhibition although the maximal glucose concentration is somewhat lower than with placebo.

Different dosing schedules were investigated in healthy volunteers in study 104. PPG data from this study
provide support for the recommendation to administer sotagliflozin immediately prior to breakfast.

Although a decrease in FPG was observed in all populations studied, statistical significance was not
achieved for all sotagliflozin treated groups in the T1DM study due to a high variability of the data.

When measured in the studies, sotagliflozin resulted in increases in GLP-1 and PPY. These findings
suggest that sotagliflozin may have additional effects apart from a reduced absorption of glucose from the
gut and an increase in the excretion of glucose from the kidney. The relevance of these mechanisms of
action, however, still remains uncertain.

In study 107, which included in total 31 T2DM patients with moderate or severe renal impairment, a
statistically significant increase in UGE compared to placebo was observed. However the UGE was lower,
with the 400 mg dose, than observed with the 300 mg dose when administered to patients with T2DM and
normal renal function. Furthermore, only a 20% reduction in post-meal plasma glucose AUC0-4h was
observed compared to a 30-35% reduction in patients with normal renal function. These data indicate
that the SGLT1 is not the only contributor to the effect on PPG. This raise concerns with regards to the use
of sotagliflozin in patients with renal impairment.

Assessment report
EMA/178275/2019 Page 39/125
The QT-study (Study 109) was of adequate design and showed that single oral doses of sotagliflozin 800
mg and 2000 mg did not cause prolongation of the QTcI interval in excess of the threshold of 10 msec.
Thus the outcome is consistent with a negative QT study according to the ICH E14 guideline.

Pharmacodynamic interactions have not been specifically discussed by the applicant. The most important
interaction is the interaction with insulin. The evaluation of this interaction is part of the objectives in the
phase 3 studies, thus this is discussed in the efficacy part of this report.

Population exposure-response (E-R) analyses were conducted to characterize the relationship between
sotagliflozin exposure and HbA1c, BW and 2-hour PPG. Adequate methods have been used in model
development and evaluation in all presented exposure-response analyses. The data indicate that there is
a large overlap in exposure for the two doses investigated and that the difference between the two doses
is very small with regards to the effects on HbA1c and body weight. The data on PPG is difficult to interpret
as inter-individual variability could not be assessed.

The population exposure-response related to safety, i.e. DKA and SH, was also assessed. There was a
trend towards a higher risk of DKA with higher exposure, but the finding is uncertain due to the low
number of events. No association between exposure and the risk of SH was observed.

2.4.5. Conclusions on clinical pharmacology

Overall the pharmacokinetics of sotagliflozin have been appropriately described.

The pharmacodynamic profile of sotagliflozin has been adequately characterised in the populations
studied. There is however very limited pharmacodynamic data in patients with renal impairment. This is
of concern, taking sotagliflozin’s mechanism of action into account. The available data in T2DM patients
with moderate renal impairment indicate an effect of sotagliflozin in patients with eGFR 45-60
mL/min/1.73 m2.

2.5. Clinical efficacy

The clinical development program included seven studies (Table 4). Studies 202 and 206 were dose
response studies. Study 203 was a proof-of concept study in T1DM patients; this study is discussed in the
pharmacology section of this report and not further discussed in this section.

Three Phase 3 studies were included in the sotagliflozin T1DM program (studies 309, 310, and 312).
Studies 309 and 310, which were conducted in North America and in Europe and Israel, respectively,
included intensive insulin optimization, and evaluated 200 mg versus 400 mg versus placebo. The third
Phase 3 study, Study 312, was conducted globally, without insulin optimization.

In the following all three studies are discussed in parallel to facilitate comparison between studies
309/310 to study 312. For a summary of the single studies, please refer to the summary of efficacy tables
(Table 20, Table 21 and Table 22)

Data from a small supportive study (204) was also provided.

2.5.1. Dose response studies

Study 202 – T2DM

This was a Phase 2, placebo-controlled, dose-ranging study conducted to assess the safety, tolerability,
and efficacy of 4 dosing regimens of sotagliflozin administered as add-on to metformin in patients with

Assessment report
EMA/178275/2019 Page 40/125
T2DM. Patients were randomly assigned (1:1:1:1:1 ratio) to 1 of 5 dosing groups: sotagliflozin 75 mg qd,
200 mg qd, 400 mg qd (2 x 200-mg tablets), or 200 mg bid, or placebo tablets.

The study consisted of a 2­week (minimum) Screening Period, 12-week Treatment Period, and a 2˗week
Follow-up Period. The primary endpoint was the change from Baseline to Week 12 in A1C; the main
secondary endpoints were the proportion of patients achieving an A1C value of <7.0% at Week 12 and
changes from Baseline in FPG, 3-hour oral glucose tolerance test (OGTT), body weight, and blood
pressure.

Eligible patients were adults, aged 18 to 75 years inclusive, with a confirmed diagnosis of T2DM, A1C
7.0% to 10.5%, on a stable dose of metformin monotherapy of ≥1500 mg/day for at least 8 weeks prior
to Screening, and for the duration of the study.

Patient disposition and demographics

A total of 299 patients were equally randomized at 52 sites in the US and were included in the
intent-to-treat (ITT) Population, which consisted of all subjects randomized to receive study drug.
Overall, 267 patients (89.3%) completed the study. The treatment groups were similar in terms of age,
ethnicity, race, BMI, and physical characteristics at Baseline. The proportion of females was higher in the
sotagliflozin 200 mg qd group (71.7%) and lower in the sotagliflozin 75 mg qd group (42.4%) compared
to the overall proportion of females (54.8%). The mean age overall was 55.9 years. The overall
proportion of white patients was 84.3% and the overall proportion of Hispanic or Latino patients was
27.4%. Baseline mean A1C ranged from 7.92 to 8.35%. The mean BMI overall was 33.08 kg/m².

Efficacy results

At Week 12, all sotagliflozin dose regimens showed a statistically significant reduction from Baseline
compared with placebo for A1C, the primary efficacy endpoint (Figure 6). The least squares (LS) mean
difference from placebo showed the greatest reduction for the 400 mg qd group (­0.79%; p <0.001),
followed by the 200 mg bid group (­0.61%; p <0.001), the 200 mg qd group (­0.34%; p = 0.018), and
the 75 mg qd group (­0.33%; p = 0.025).

Figure 6 - Study 202: Mean Change in A1C (%) from Baseline to Week 12

The proportion of patients who achieved an A1C <7.0% at Week 12 was highest in the 400 mg qd group
(38.6%), and this proportion was significantly different from the placebo group (24.6%; p = 0.022). The
difference from placebo was also statistically significant in the 200 mg bid group (29.8%; p = 0.036).

Assessment report
EMA/178275/2019 Page 41/125
The 3 higher sotagliflozin dose regimens (200 mg, 200 mg BID and 400 mg) were statistically significant
compared with placebo in reducing FPG, in reducing blood glucose area under the curve (AUC) during the
3-hour oral glucose tolerance test (all p ≤0.005) and in reducing body weight (all p ≤0.001).

Study 206

This was a Phase 2b, multicenter, randomized, double-blind, placebo-controlled, parallel-group trial to
evaluate the efficacy and safety of several doses of sotagliflozin in patients with T1DM with inadequate
glycaemic control. The study consisted of a 2-week Screening Period, 2-week Single-blind Placebo Run-in
Period, 12-week Double-blind Treatment Period, and a 30­day (post last dose) follow-up phone call.
During the 12-week Double-blind Treatment Period, the protocol stipulated that the patients’ insulin dose
should remain stable (within ±20% of the Baseline total insulin dose), unless an adjustment in insulin
dose was required for recurrent hypoglycaemia. The primary efficacy endpoint was the change in A1C
from Baseline to Week 12. The secondary efficacy endpoints included change from Baseline to Week 12
for each of the following: 2-hour PPG following a standardized Mixed Meal; body weight (absolute and
percent change); UGE; and FPG.

Eligible patients were 18 years and older with confirmed diagnosis of T1DM made at least 1 year prior to
informed consent, a screening A1C from 7.0% to 10.0% (inclusive), and a basal insulin that had not
changed ≥20% for the 2 weeks prior to the Screening Visit. Patients were being treated with insulin(s) or
insulin analogue(s) delivered via CSII or MDI; using the same method of insulin delivery for the 3 months
prior to the Screening Visit. On Day 1, of the Double-blind Treatment Period, patients were randomly
assigned (1:1:1:1 ratio) to sotagliflozin (75 mg, 200 mg, or 400 mg [given as 2 x 200-mg tablets]) or
matching placebo, all given once daily before the first meal of the day as an adjunct to their insulin
therapy.

Patient disposition and demographics

A total of 141 patients were equally randomized at 17 study sites in the US, and included in the mITT
population. By randomization stratum, there were a total of 73 (51.8%) patients who received insulin by
CSII and 68 (48.2%) patients who received insulin by MDI. A total of 130 patients completed the study.
At Baseline, the mean (SD) age was 45.6 (13.29) years. The majority of patients were white
(131, 92.9%). Sex was acceptably distributed given the group size (male 48.2%; female 51.8%). The
mean (SD) Baseline body weight was 85.48 (18.557) kg, and the mean (SD) BMI was
29.17 (5.569) kg/m2. The mean A1C at Baseline was comparable across the sotagliflozin groups.

Efficacy results

The primary efficacy objective was met by observing a dose relationship from Baseline to Week 12 in A1C
reduction across placebo and the 3 sotagliflozin treatment groups (p = 0.008) (Figure 7).

Assessment report
EMA/178275/2019 Page 42/125
Figure 7 - Study 206: mean change in A1C from Baseline to Week 12

The outcome for the secondary endpoints (change from Baseline to Week 12 for 2-hour PPG; body weight
and UGE) are shown in Figure 8.

A dose-dependent effect of sotagliflozin on mean reduction from Baseline in 2-hour PPG following a
standardized mixed meal at Week 12 was observed (linear trend p = 0.008). The LS mean difference from
placebo in change from Baseline in the 2-hour PPG at Week 12 was ­20.3 mg/dL in the sotagliflozin 75 mg
group (p = 0.27), ­27.4 mg/dL in the sotagliflozin 200 mg group (p = 0.15), and ­49.5 mg/dL in the
sotagliflozin 400 mg group (p = 0.006).

There was a statistically significant mean decrease from Baseline to Week 12 in absolute body weight
compared with placebo for all groups, with the greatest effect observed for sotagliflozin 200 mg and 400
mg groups (p <0.001). The LS mean difference from placebo in change from Baseline in absolute body
weight was ­2.37 kg in the sotagliflozin 200 mg group and ­2.61 kg in the sotagliflozin 400 mg group.

The LS mean difference from placebo in change from Baseline in UGE at Week 12 was statistically
significant for all sotagliflozin dose groups: 41.76 g/day in the sotagliflozin 75 mg group (p = 0.006),
57.73 g/day in the sotagliflozin 200 mg group (p <0.001), and 70.45 g/day in the sotagliflozin 400 mg
group (p <0.001).

Figure 8 - Study 206: secondary and other efficacy endpoints

The LS mean difference from placebo in change from Baseline in FPG at Week 12 was -8.6 mg/dL in the
sotagliflozin 75 mg group, ­8.9 mg/dL in the sotagliflozin 200 mg group, and ­21.4 mg/dL in the
sotagliflozin 400 mg; none of these decreases were statistically significant compared with placebo.

Assessment report
EMA/178275/2019 Page 43/125
2.5.2. Main studies

Methods

Studies 309 and 310

Studies 309 and 310 had the same design. Both studies were Phase 3, multicenter, randomized,
double-blind, placebo-controlled, parallel-group studies to evaluate the efficacy, safety, and tolerability of
sotagliflozin 200 mg and 400 mg versus placebo administered qd as an adjunct therapy in adults with
T1DM who had inadequate glycaemic control with insulin therapy (delivered via continuous subcutaneous
insulin infusion [CSII] or multiple daily injection [MDI]). For each study, the total duration of exposure to
study drug was up to 52 weeks. A key design feature for studies 309 and 310 was insulin optimization.

The primary objective was to establish the superiority of sotagliflozin 200 mg and/or 400 mg administered
before the first meal of the day compared with placebo, as measured by the change in A1C from Baseline
to Week 24.

The total duration of study participation was up to 64 weeks. The study schema is presented in Figure 9.

Figure 9 - Studies 309 and 310: Study Schema

* Patients who participated in the optional DEXA substudy were to complete the Baseline DEXA -2weeks from
the Day 1 visit. The visit window for DEXA after the Day 1 visit was to be ± 2 weeks

** Patients who participated in the optional CGM substudy, and all patients who were screened after IRB approval of
Amendment 2, were to complete the Week 53 Follow-up visit

Two substudies were included in both 309 and 310: a CGM substudy, designed to obtain continuous data
on blood glucose values and a dual-energy X-ray absorptiometry (DEXA) substudy, designed to evaluate
fat mass and bone density. Patients could have participated in 1 or both substudies.

Patients eligible were 18 years and older with a diagnosis of T1DM made at least 1 year prior to informed
consent, a Screening A1C of 7.0% to 11.0% (inclusive), an eGFR >45 mL/min/1.73 m2, and a triglyceride
value <600 mg/dL. Patients were being treated with insulin(s) or insulin analogue(s) delivered via CSII or

Assessment report
EMA/178275/2019 Page 44/125
MDI; using the same method of insulin delivery for the 3 months prior to the Screening Visit. Antidiabetic
treatments other than insulin were prohibited. Patients with a history of SH or DKA within 1 month prior
to Screening were excluded.

All patients who completed the Core Treatment Period were eligible to participate in the LTE Period.

Patients could participate in 1 or both (CGM and DEXA) substudies. Patients selected for DEXA evaluation
had to satisfy additional eligibility criteria: age 30 to 80 years old, inclusive and not on any medications
that could affect bone density for at least 3 years.

Key design features related to Insulin Optimization in Studies 309 and 310

In order to evaluate the efficacy of sotagliflozin beyond what can be provided by insulin alone, all patients
in Studies 309 and 310 entered a 6-week insulin optimization period prior to randomization, with the
objective of improving glycaemic control using insulin alone. Patients were maintained on optimized
insulin after being randomized to 1 of 2 doses of sotagliflozin (200 mg or 400 mg) or placebo. During
insulin optimization, insulin adjustment was assessed by an insulin dose monitoring committee (IDMC) of
independent experts from Week -5 to Week 24 (time of primary and secondary endpoint assessment). To
be eligible for inclusion in both studies, the screening (pre-optimization) A1C was required to be ≥7.0%.
This A1C inclusion criterion was not repeated after optimization, therefore the study population included
approximately 20% patients with A1C <7.0% at Baseline.

Study 312

Study 312 was a Phase 3, multicenter, randomized, double-blind, placebo-controlled, parallel-group

study to evaluate the net benefit (defined as the proportion of patients with A1C <7.0% and no episode
of SH and no episode of DKA from Randomization to Week 24) of sotagliflozin 400 mg versus placebo
administered qd as an adjunct to insulin in patients with T1DM. This study did not utilize insulin
optimization, and treatment was not confined to any specific insulin regimen.

The primary objective of this study was to demonstrate the superiority of sotagliflozin 400 mg qd
administered before the first meal of the day compared with placebo in the proportion of patients with
A1C <7.0% at Week 24 and no episode of SH and no episode of DKA after Randomization.

The total duration of study participation was up to 32 weeks. The study schema is presented in Figure 10.

Assessment report
EMA/178275/2019 Page 45/125
Figure 10 - Study 312: Study Schema

Sotagliflozin 400 mg qd dosage was used in Study 312 to evaluate safety exposure at the higher dose.
The total duration of exposure to study drug was up to 24 weeks.

Eligible patients were ≥18 years of age, with a BMI ≥18.5 kg/m2, a confirmed diagnosis of T1DM made at
least 1 year before Screening, and a Screening A1C of 7.0% to 11.0%, (inclusive), an eGFR >45
mL/min/1.73 m2, and a triglyceride value <600 mg/dL. Use of any antidiabetic agent other than insulin or
insulin analogue within 8 weeks prior to Screening was excluded. Patients using any background insulin
regimen including CSII, MDI, intermediate acting, and pre-mixed insulins, were eligible to participate.
Patients with a history of SH or DKA within 1 month prior to Screening were excluded.

Glycaemic goals and adjustment of insulin dose in studies 309, 310 and 312

The recommended target for fasting/preprandial glucose was 80-130 mg/dL (4.4 to 7.2 mmol/L).
Recommended target for 1-2-hour postprandial glucose (PPG) by SMBG was <180 mg/dL
(<10.0 mmol/L). Goals could have been adapted for individual patients based on investigator discretion.

For the first meal on Day 1, patients were to be instructed to decrease their calculated (or usual) mealtime
carbohydrate bolus insulin by 30%. Subsequent adjustments in insulin dosing were to be made by the
Investigator, based on SMBG trends. The mealtime insulin adjustment recommendations for Day 1 were
based on results from Study 203. It was expected that the insulin-to-carbohydrate ratio would be
decreased by approximately 30% with sotagliflozin treatment. It was also predicted that such an
adjustment would convey appropriate margins of safety for both the treated and placebo groups. The
insulin-to-carbohydrate was to be subsequently adjusted as frequently as required by the Investigator to
meet glycaemic goals recommended for this study. No change in high blood glucose correction factor
(sliding scale), basal (or nonbolus) insulin, or basal rate was to be recommended at initiation of study
drug, although the Investigator was not to be prohibited from making any changes in any insulin dose
consistent with meeting glycaemic goals recommended for this study.

Assessment report
EMA/178275/2019 Page 46/125
Outcomes/endpoints

Table 6 - Main efficacy endpoints in phase 3 studies

Studies 309 and 310 Study 312

Primary Endpoint A1C change from Baseline at Week 24 Net benefit at Week 24a
Secondary Net benefit at Week 240 A1C change from Baseline at Week 24
Endpoints
(in hierarchical
order)
Body Weight change from Baseline at Week Body Weight change from Baseline at Week
24 24
Bolus insulin change from Baseline at Week 24 SBP change from Baseline at Week 16 (in
patients with SBP ≥130 mm Hg at Baseline)
FPG change from Baseline at Week 24 Bolus insulin change from Baseline at Week
24
DTSQs score change from Baseline at Week
24
DDS2 score change from Baseline at Week 24
Other Endpoints SBP change from Baseline at Week 12 (in FPG change from Baseline at Week 24
patients with SBP ≥130 mm Hg at Baseline)
Composite Endpoints evaluating benefit/risk: Composite Endpoints evaluating
Other clinical benefits (achieving benefit, benefit/risk: Other clinical benefits
defined as A1C reduction, while having low risk (achieving benefit, defined as A1C
in other parameters such as SH, DKA, body reduction, while having low risk in other
weight, and insulin dosing). parameters such as SH, DKA, body weight,
and insulin dosing).
Net benefit was defined as the proportion with A1C <7.0% at week 24 and no episode of SH and no episode of DKA after randomization.
In studies 309 and 310, SBP changes were evaluated at Week 12, consistent with the protocol design to
keep antihypertensive medication doses stable until Week 12. Additionally, the mean change from
Baseline in SBP was evaluated in a subset of patients with Baseline SBP ≥130 mm Hg at Weeks 12 and 24.

Table 7 CGM primary, key secondary and other endpoints (mITT CGM population)

Pooled 309 and 310

Primary Endpoint Percent time spent in target range (70 to 180 mg/dL)
Secondary % time spent outside the target range, AUC outside the target range (by hyperglycaemia and
Endpoints hypoglycaemia), % time and AUC in hyperglycaemia or hypoglycaemia by Sleep-Wake Time
Blocks
Average minutes/day in target, hyperglycaemia or hypoglycaemia range, number of hypoglycaemic
events, MAGE, MAG, coefficient of variation, mean SD, mean daily glucose, and distance travelled
(all endpoints Week 24)

Table 8 DEXA primary, key secondary and other endpoints (mITT DEXA population)

Pooled 309 and 310

Primary Endpoint Total fat mass (Week 24)
Secondary Total lean mass and % total fat (Week 24, 52)
Endpoints

Assessment report
EMA/178275/2019 Page 47/125
Sample size

Studies 309 and 310

Separately for both Studies 309 and 310, the sample size estimate was based on satisfying design
assumptions and statistical testing requirements for the primary efficacy endpoint. The final sample size
was estimated as 250 randomized patients per treatment group (750 total) for each study. The two
substudies, a CGM substudy and a DEXA substudy, will target approximately 70 patients per treatment
group to satisfy their primary objectives.

Study 312

For Study 312, the sample size was based on satisfying assumptions made for the primary efficacy
endpoint of net clinical benefit and to provide a suitable number of patients so that a reliable estimate of
treatment group differences in SH could be made. Based on these considerations, 700 randomized
patients were required per treatment group or 1400 total patients.

Randomisation

The desired balances were to be accomplished by use of randomly permuted blocks of fixed size. An
Interactive Voice/Web Response System (IXRS) was to be used as a central mechanism to assign patients
to study treatment. Patients who terminated the study early were not to be replaced in this study.

Studies 309 and 310

Patients were to be randomly assigned among 3 parallel treatment groups in a 1:1:1 manner. The
treatment randomization schedule, was centralized and stratified by insulin delivery method (MDI or
CSII) and Week -2 A1C (≤8.5%, >8.5%). Enrolment of patients using insulin via MDI was to be capped
at 60%; no more than 150 patients on MDI were to be randomly assigned to each treatment arm. Upon
reaching this threshold, patients were to have been randomly assigned to the treatment groups, being
stratified only by Week -2 A1C (≤8.5%, >8.5%).

Study 312

Patients were to be randomly assigned between 2 treatment groups in a parallel, 1:1 manner. The
treatment randomization schedule was centralized and stratified by BMI at Screening (<25 kg/m2, ≥25
kg/m2), Week 2 A1C (≤9.0%, >9.0%), and use of CSII at Screening (yes, no).

Blinding (masking)

In all the double-blind studies, the designated group and treatment assigned to each patient was not to
have been revealed to the Investigator, site staff, the patient, or the Sponsor, or designee, until the
decision was made to unblind the study. To maintain the double-blind design, A1C, FPG, and 2­hour PPG
results (in the CGM substudy) were masked from the visit after Randomization to the visit before
Week 24. In addition, because of the glycosuria effect of sotagliflozin, urine glucose values were to have
been masked to all study staff.

Statistical methods

Studies 309 and 310

For summaries based on use of inferential statistics, all significance tests were 2-sided and used a 0.05
α-level. Confidence limits (CLs) were likewise 2-sided and used a 95% confidence coefficient.

Assessment report
EMA/178275/2019 Page 48/125
All efficacy analyses were performed on the modified intent to treat (mITT) population. A supportive
analysis of the primary endpoint was also performed on the per-protocol (PP) population. All safety
analyses were performed on the Safety Population.

Randomized Population: All randomly assigned patients were included in the Randomized Population.
Randomized patients were analysed according to their randomized treatment.

Safety Population: All randomly assigned patients treated with at least 1 dose of study drug were included
in the Safety Population. Safety patients were analysed according to their actual treatment received on
Day 1.

Modified Intent-to-treat Population: The mITT population consists of all randomly assigned patients who
had taken at least 1 dose of study drug. Patients in the mITT population were analysed according to their
randomized treatment.

Per-Protocol Population: The PP Population comprises all patients in the mITT population who completed
treatment through the primary CT Period of 24 weeks, and did not have any significant protocol deviations
during this time. Significant protocol deviations were defined as those deviations considered having a
major effect on the collection or interpretation of the primary efficacy endpoint. In addition, a separate PP
Population was defined for the CGM substudy for analysis of the change from Baseline to Week 24 in the
2-hour PPG following the standardized Mixed Meal.

The primary efficacy endpoint was the change from Baseline to Week 24 in A1C in either sotagliflozin dose
(200 mg or 400 mg) compared with placebo. The primary analysis of the primary efficacy endpoint used
mixed-effects MMRM statistics based on the restricted maximum likelihood method for estimation and to
be performed using the mITT Population. The analysis model included fixed, categorical effects of
treatment, randomization strata of insulin delivery method (MDI, CSII), randomization strata of Week -2
A1C (≤8.5%, >8.5%), time (study week), a treatment-by-time interaction, and Baseline A1C-by-time
interaction as a covariate. An unstructured (co)variance structure was used to model the within-patient
errors. The model adjusted mean (i.e. least squares [LS] mean) change in A1C from Baseline to Week 24
for each treatment group was estimated in the framework of this model, as well as the between group
differences (comparing sotagliflozin to placebo) and the 95% CLs for the difference.

In addition, the primary efficacy endpoint was analysed for the PP Population using an analysis of
covariance (ANCOVA) model fitted for the fixed, categorical effects of treatment, randomization strata of
insulin delivery method, randomization strata of Week -2 A1C, and Baseline A1C as a covariate.

A key assumption for drawing valid conclusions using the MMRM analysis was that the reason for missing
data was a function of the MAR mechanism. Since one could not be fully certain that other mechanisms
might underlie the reason for missing data (e.g., MNAR), it was important to perform sensitivity analyses
of the MMRM results. Under an assumption of MNAR, the PMM with control based pattern imputation was
used in the sensitivity analysis for this study. The model’s actual implementation dependent on the extent
and pattern of missing data and, as such, it might not be used. For the analysis of binary efficacy
endpoints, missing observations at Week 24 were imputed as nonresponse (i.e., nonresponder
imputation).

The secondary efficacy endpoints were measured for either sotagliflozin dose compared with placebo for
each of the following:

• Proportion of patients with A1C <7.0% (at Week 24) and no episode of severe hypoglycaemia,
and no episode of DKA (severe hypoglycaemia and DKA occurrence over the cumulative
randomized Double-blind 24-week CT Period)

• Change from Baseline in body weight at Week 24 (absolute and percent change)

Assessment report
EMA/178275/2019 Page 49/125
 • Change from Baseline in mean daily bolus insulin dose at Week 24

• Change from Baseline in FPG at Week 24

• Change from Baseline in Diabetes treatment satisfaction as measured by DTSQs scores at Week
24

• Change from Baseline in Diabetes Distress as measured by DDS2 scores at Week 24

For continuous secondary efficacy endpoints, an MMRM model, as specified for the primary efficacy
variable was applied, with the corresponding endpoint and Baseline value-by-time covariate specific for
that secondary endpoint used in the model. Since the continuous secondary efficacy endpoints are all
changes (or percent changes, for some) from Baseline to Week 24, all post-Baseline observations
collected during the CT Period, and following application of the visit window rules, were used as
dependent variables in the MMRM.

For the binary secondary efficacy endpoint, a Cochran-Mantel-Haenszel (CMH) test stratified by the
different levels of the randomization stratification factors of insulin delivery method (MDI, CSII) and Week
-2 A1C (≤8.5%, >8.5%) was used.

Multiplicity in statistical testing of the efficacy variables occurred from 2 main sources: (a) testing of the
primary endpoint and multiple secondary endpoints, and (b) testing of 2 sotagliflozin dose groups against
placebo for each endpoint. These considerations yielded 14 hypotheses to be tested that can be grouped
into 7 families: each family corresponded to the specific endpoint under test.

Family F1 consisted of the sotagliflozin 200 mg versus placebo and sotagliflozin 400 mg versus placebo
comparisons for the primary endpoint. Family F2 included the same treatment group comparisons for the
first listed secondary endpoint, F3 included the same comparisons for the second listed secondary
endpoint, and so on. The 7 families were to be tested sequentially with the restriction that the test of each
treatment group comparison required all prior tests of that particular comparison to meet statistical
significance criteria. The primary endpoint hypotheses were to be tested by a Bonferroni procedure with
α = 0.05 (2-sided) and use of equal weights so that the per comparison error rate = 0.025 (2-sided). The
raw p-value for each treatment contrast was compared with α = 0.025 and if the raw p-value was less
than or equal to 0.025, the comparison was declared statistically significant and testing for that contrast
could proceed to the next listed endpoint. Consistent with testing the primary endpoint, the family-wise
error rate (FWER) within each secondary endpoint family was 0.05 (2-sided), and with hypothesis weights
of 0.5 assigned to each contrast within each testable family, the per comparison α-level = 0.025
(2-sided). Formal testing was to stop at that endpoint for which a raw p-value exceeded 0.025.

Progression in testing across the hypothesis families was to be carried out, in essence, using a tree
gatekeeping test procedure so that the study-wise error rate across all primary and secondary
hypotheses tested was strongly controlled at α = 0.05.

CGM study: Due to recruitment of the CGM substudy not meeting the planned target, CGM data from the
309 study was to be pooled with data from Study LX4211.310. This eventuality was prespecified in the
protocols for the respective studies prior to unblinding any CGM data.

Study 312

The primary efficacy endpoint was the proportion of patients with A1C <7.0% at Week 24 and no episode
of severe hypoglycaemia and no episode of DKA from Randomization to Week 24. The analysis of this
endpoint was based on the mITT population. The frequency and percentage of patients with A1C <7.0%
at Week 24 and no episode of severe hypoglycaemia or episode of DKA from Randomization to Week 24
were presented by treatment group.

Assessment report
EMA/178275/2019 Page 50/125
Analysis of the primary efficacy endpoint at Week 24 was performed using the CMH test and 95% CLs
based on normal approximation methods with a continuity correction factor, stratified by the different
levels of the randomization stratification factors of BMI at Screening (<25 kg/m2, ≥25 kg/m2), Week -2
A1C (≤9.0%, >9.0%), and use of CSII at Screening (yes, no). The treatment group comparisons were
performed at Week 24 only. The individual components of the endpoint were summarized separately
using descriptive methods.

Analysis of the secondary efficacy endpoints were based on the mITT population. The secondary efficacy
endpoints were measured as change from Baseline in sotagliflozin 400 mg compared with placebo and
were analysed using MMRM model using the restricted maximum likelihood method for estimation. The
analysis mode included treatment, randomization strata of BMI at Screening (<25 kg/m2, ≥25 kg/m2),
randomization strata of Week -2 A1C (≤9.0%, >9.0%), randomization strata of use of CSII at Screening
(yes, no), time (study week), and a treatment-by-time interaction as fixed categorical effects, and
baseline-dependent variable-by-time interaction as a covariate. For the endpoint of percent change from
Baseline, the analysis model was not to include the interaction covariate. An unstructured (co)variance
structure was used to model the within-patient errors.

The test for superiority of sotagliflozin versus placebo based on the primary efficacy endpoint was
performed at the 2-sided 0.05 α-level. If this null hypothesis was rejected, analysis of the secondary
endpoints at Week 24 were to occur in a sequential, prespecified hierarchy so that the overall type I error
rate was strongly controlled at a 2-sided, 0.05 α-level across these endpoints. The order of testing for the
secondary endpoints was 1) A1C change from Baseline at Week 24, 2) Body weight at Week 24 change
from Baseline (absolute and percent change; the absolute change will be used in the sequence of
analyses), 3) SBP change from Baseline at Week 16 in the subset of patients with Baseline SBP ≥130 mm
Hg, 4) Percent change from Baseline in bolus insulin dose at Week 24.

Missing data was handled as described for studies 309 and 310.

Pool E1 and E2

Efficacy analyses from all Phase 3 studies were pooled into 2 groupings, ES1 and ES2. The ES1 pool
included the 52-week Studies 309 and 310, both of which defined change from Baseline in A1C at Week
24 as their primary endpoint. The ES2 pool included placebo and the 400 mg sotagliflozin dose groups
from all 3 Phase 3 studies. Both pools were designed to allow for analysis of efficacy endpoints across
several subgroups of clinical interest. Another objective for the ES1 pool was to provide increased
precision for the CGM and DEXA substudy assessments.

The pooled analyses of the primary and secondary endpoints followed the methods provided in the
individual study SAPs. No multiple testing adjustments were performed for the pooled analyses.

Results

In the following, the results for studies 309, 310 and 312 are discussed in parallel. For the outcome of the
studies one by one, please refer to the efficacy tables (Table 20, Table 21 and Table 22).

Assessment report
EMA/178275/2019 Page 51/125
Participant flow

Figure 11 Study 309: Flow Diagram of the Study – All Patients Enrolled

Abbreviations: A1C = haemoglobin A1C; CGM = continuous glucose monitoring; CT = Core Treatment;
DEXA = dual-energy X-ray absorptiometry; mITT = modified intent-to-treat; MMTT = Mixed Meal tolerance test; n =
number of patients; PP = per-protocol; PPG = postprandial glucose.
Note: Patients may have had more than 1 exclusionary event.

Assessment report
EMA/178275/2019 Page 52/125
Figure 12 Study 310: Flow Diagram of the Study – All Patients Enrolled

Abbreviations: A1C = haemoglobin A1C; CGM = continuous glucose monitoring; CT = core treatment; DEXA =
dual-energy X-ray absorptiometry; mITT = modified intent-to-treat; MMTT = mixed meal tolerance test; n = number
of patients; PP=per-protocol; PPG = postprandial glucose.
Note: Patients may have had more than 1 exclusionary event.

Assessment report
EMA/178275/2019 Page 53/125
Figure 13 Study 312: Flow Diagram of the Study – All Patients Enrolled

Abbreviations: A1C = haemoglobin A1C; mITT = modified Intent-to-Treat; n = number of patients; PP = per-protocol.
Patients may have been excluded or discontinued for more than 1 reason.
Patients may have been excluded for more than 1 reason.

Conduct of the studies

In study 310 there was 1 incident where ethical conduct of the study was compromised. Inspections have
been performed by the Lithuanian health authority, which concluded that the safety of the subjects and
integrity of their data were not compromised, as the initial IEC approvals had been obtained.

Assessment report
EMA/178275/2019 Page 54/125
All patients in Study 310 were to have trough plasma levels of sotagliflozin and its main metabolite
determined at Weeks 12 and 24. The Sponsor identified 42 of the 782 randomized patients who had
plasma PK samples suggesting that they had received incorrect study drug or had failed to take study
drug as required for some periods during the study. The Sponsor performed a root cause investigation. As
a result of this investigation, the Sponsor believes that the issues were a consequence of specific site
errors in dispensing and handling of investigational product, a laboratory sample handling error, as well
as patient noncompliance, and affect only limited number of patients (42 patients), 5% of the total
patients randomized in the 310 study. Because of these PK sample findings, additional analyses were
performed to assess their potential impact on overall study results. Sensitivity analyses applying several
statistical methods were performed on the safety and efficacy data to evaluate the potential impact of the
anomalous findings on study outcomes. Results of these analyses showed no impact on the primary and
secondary efficacy nor safety endpoints, nor on the benefit-risk profile of sotagliflozin as demonstrated in
this study.

Baseline data

Table 9 Summary of demographics for Studies 309, 310, and 312 (mITT population)

Study 309 Study 310 Study 312

Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263) (N = 703) (N = 699)

Mean (SD) age at study entry 45.2 46.6 46.4 39.7 42.3 41.7 42.4 43.3
(years) (12.72) (13.48) (13.12) (13.42) (13.59) (13.23) (14.04) (14.17)
Age at study entry (years), n
(%)
≥18 to <26 20 (7.5) 21 (8.0) 19 (7.3) 43 (16.7) 39 (14.9) 39 (14.8) 98 (13.9) 90 (12.9)
≥26 to <50 140 (52.2) 128 (48.7) 132 (50.4) 151 (58.5) 138 (52.9) 149 (56.7) 378 (53.8) 371 (53.1)
≥50 to <65 89 (33.2) 91 (34.6) 89 (34.0) 50 (19.4) 75 (28.7) 65 (24.7) 181 (25.7) 183 (26.2)
≥65 to <75 18 (6.7) 20 (7.6) 21 (8.0) 13 (5.0) 8 (3.1) 8 (3.0) 45 (6.4) 50 (7.2)
≥75 1 (0.4) 3 (1.1) 1 (0.4) 1 (0.4) 1 (0.4) 2 (0.8) 1 (0.1) 5 (0.7)
Age at T1DM diagnosis (years)
Mean (SD) 21.0 21.6 22.4 21.7 24.0 22.8 22.8 22.8
(12.88) (13.37) (13.14) (12.59) (12.88) (12.54) (13.51) (13.65)
Age at T1DM diagnosis (years)
<18 119 (44.4) 120 (45.6) 112 (42.7) 111 (43.0) 88 (33.7) 99 (37.6) 296 (42.1) 294 (42.1)
≥18 149 (55.6) 143 (54.4) 150 (57.3) 147 (57.0) 173 (66.3) 164 (62.4) 407 (57.9) 405 (57.9)
Sex
Male 137 (51.1) 126 (47.9) 120 (45.8) 134 (51.9) 139 (53.3) 133 (50.6) 339 (48.2) 358 (51.2)
Female 131 (48.9) 137 (52.1) 142 (54.2) 124 (48.1) 122 (46.7) 130 (49.4) 364 (51.8) 341 (48.8)
Race
White 244 (91.0) 241 (91.6) 246 (93.9) 250 (96.9) 252 (96.6) 250 (95.1) 621 (88.3) 619 (88.6)
Black 9 (3.4) 11 (4.2) 8 (3.1) 1 (0.4) 0 0 22 (3.1) 24 (3.4)
Weight (kg)
Mean (SD) 87.30 86.96 86.50 81.08 81.93 81.97 81.55 82.40
(17.709) (18.539) (18.004) (16.857) (17.386) (17.963) (17.032) (17.131)
BMI (kg/m2)
Mean (SD) 29.55 29.81 29.63 27.50 27.97 27.85 28.10 28.29
(5.188) (5.686) (5.297) (5.170) (5.275) (4.921) (5.183) (5.128)
BMI (kg/m2), category
<18.5 0 0 0 0 3 (1.1) 2 (0.8) 1 (0.1) 3 (0.4)
≥18.5 to <25 53 (19.8) 60 (22.8) 51 (19.5) 88 (34.1) 79 (30.3) 74 (28.1) 205 (29.2) 201 (28.8)
≥25 to <30 101 (37.7) 82 (31.2) 97 (37.0) 98 (38.0) 95 (36.4) 109 (41.4) 279 (39.7) 259 (37.1)
≥30 114 (42.5) 121 (46.0) 114 (43.5) 72 (27.9) 84 (32.2) 78 (29.7) 218 (31.0) 236 (33.8)
Geographic Region

Assessment report
EMA/178275/2019 Page 55/125
 Study 309 Study 310 Study 312
Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263) (N = 703) (N = 699)

North America (US and 268 263 262 0 0 0 302 (43.0) 277 (39.6)
Canada) (100.0) (100.0) (100.0)
Outside North America 0 0 0 258 261 263 401 (57.0) 422 (60.4)
(100.0) (100.0) (100.0)
Duration of T1DM (years)
<20 107 (39.9) 99 (37.6) 104 (39.7) 158 (61.2) 160 (61.3) 160 (60.8) 385 (54.8) 373 (53.4)
≥20 to <40 125 (46.6) 122 (46.4) 126 (48.1) 85 (32.9) 89 (34.1) 88 (33.5) 270 (38.4) 268 (38.3)
≥40 36 (13.4) 42 (16.0) 32 (12.2) 15 (5.8) 12 (4.6) 15 (5.7) 48 (6.8) 58 (8.3)
Insulin Delivery Method
CSII 160 (59.7) 156 (59.3) 157 (59.9) 66 (25.6) 68 (26.1) 67 (25.5) 280 (39.8) 275 (39.3)
Non-CSII 108 (40.3) 107 (40.7) 105 (40.1) 192 (74.4) 193 (73.9) 196 (74.5) 423 (60.2) 424 (60.7)

Table 10 Summary of Baseline characteristics for Studies 309, 310, and 312 (mITT
population)

Study 309 Study 310 Study 312

Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263) (N = 703) (N = 699)

A1C (%) at Week -2, n (%)

≤8.5% 228 (85.1) 226 (85.9) 227 (86.6) 202 (78.3) 205 (78.5) 208 (79.1) 417 (59.3) 423 (60.5)
>8.5% 40 (14.9) 37 (14.1) 35 (13.4) 56 (21.7) 56 (21.5) 55 (20.9) 284 (40.4) 276 (39.5)
Mean (SD) A1C (%) 7.54 7.61 7.56 7.79 7.74 7.71 8.21 8.26
(0.712) (0.735) (0.724) (0.881) (0.806) (0.819) (0.921) (0.965)
Mean (SD) A1C (mmol/mol) 58.92 59.67 59.12 61.64 61.11 60.82 66.23 66.72
(7.805) (7.984) (7.911) (9.627) (8.774) (8.945) (10.073) (10.521)
Baseline eGFR
(ml/min/1.73m2), n (%)
<60 15 (5.6) 14 (5.3) 16 (6.1) 9 (3.5) 8 (3.1) 9 (3.4) 42 (6.0) 32 (4.6)
≥60 to <90 141 (52.6) 146 (55.5) 143 (54.6) 104 (40.3) 124 (47.5) 116 (44.1) 300 (42.7) 312 (44.6)
≥90 112 (41.8) 103 (39.2) 103 (39.3) 145 (56.2) 129 (49.4) 138 (52.5) 361 (51.4) 355 (50.8)
Mean (SD) FPG (mg/dL) 153.66 155.06 148.24 160.45 163.66 165.46 163.42 165.10
(64.534) (68.710) (62.949) (65.356) (74.482) (71.059) (69.083) (71.603)
na 42 42 45 39 43 45 NA NA
Mean (SD) 2-hour PPG 227.60 198.40 202.98 232.10 223.09 216.33 NA NA
(mg/dL) (91.627) (91.819) (69.113) (107.304) (103.539) (99.922)
Mean (SD) Total Daily Insulin 0.74 0.72 0.72 0.75 0.73 0.74 0.71 0.69
Dose (IU/kg) (0.357) (0.386) (0.335) (0.295) (0.277) (0.267) (0.291) (0.276)
Mean (SD) Total Daily Insulin 66.79 65.11 64.15 61.85 60.30 61.38 58.35 56.88
Dose (IU/day) (41.265) (42.698) (37.636) (30.862) (28.963) (28.653) (29.085) (27.601)
Mean (SD) Ratio of Daily Bolus 0.46 0.45 0.46 0.50 0.51 0.51 0.48 0.47
vs. Total Insulin Doses (IU/IU) (0.128) (0.138) (0.129) (0.132) (0.134) (0.125) (0.151) (0.152)
Mean (SD) Sitting SBP (mm 120.9 120.0 119.5 123.1 123.0 123.1 121.8 122.0
Hg) (13.47) (14.84) (14.73) (15.53) (15.08) (13.69) (14.82) (15.25)
Baseline Sitting SBP (mm Hg),
n (%)
<130 204 (76.1) 203 (77.2) 202 (77.1) 173 (67.1) 175 (67.0) 183 (69.6) 499 (71.0) 496 (71.0)
≥130 64 (23.9) 60 (22.8) 60 (22.9) 85 (32.9) 86 (33.0) 80 (30.4) 203 (28.9) 203 (29.0)
<140 248 (92.5) 237 (90.1) 241 (92.0) 228 (88.4) 228 (87.4) 235 (89.4) 618 (87.9) 601 (86.0)
≥140 20 (7.5) 26 (9.9) 21 (8.0) 30 (11.6) 33 (12.6) 28 (10.6) 84 (11.9) 98 (14.0)
Baseline
Mean (SD) Sitting DBP (mm 76.4 (8.24) 76.4 (9.28) 75.3 (9.17) 76.3 (8.48) 77.4 (9.83) 76.2 (8.37) 76.7 (9.06) 76.4 (8.77)
Hg)
Baseline Pulse Rate 73.5 74.3 72.4 75.5 74.6 74.3 75.2 74.9
(beats/minute) (10.88) (10.85) (11.60) (10.87) (11.59) (11.05) (11.17) (11.54)

Assessment report
EMA/178275/2019 Page 56/125
 Study 309 Study 310 Study 312
Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263) (N = 703) (N = 699)
a 2hr-postprandial glucose was evaluated only in the CGM population in Studies 309 and 310.

Numbers analysed

Table 11 - Summary of patient disposition for Studies 309, 310, and 312 (randomized
population)

Study 309 Study 310 Study 312

Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263) (N = 705) (N = 700)
n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%)
Randomized 268 263 262 258 261 263 705 700
(100.0) (100.0) (100.0) (100.0) (100.0) (100.0) (100.0) (100.0)
mITT Population 268 263 262 258 261 263 703 699
(100.0) (100.0) (100.0) (100.0) (100.0) (100.0) (99.7) (99.9)
Completed 24 Weeks 235 240 236 236 239 240 624 605
(87.7) (91.3) (90.1) (91.5) (91.6) (91.3) (88.5) (86.4)
Completed 52 Weeks 218 228 221 225 226 227 NA NA
(81.3) (86.7) (84.4) (87.2) (86.6) (86.3)

Table 12 - Summary of patient disposition for the ES1 pool and the ES2 pool (randomized
population)
a
ES1 Pool ES2 Pool
Placebo SOTA SOTA All SOTA Placebo SOTA
200 mg 400 mg (N = 1049) 400 mg
(N = 526) (N = 524) (N = 525) n (%) (N = 1231) (N = 1225)
n (%) n (%) n (%) n (%) n (%)
Randomized 526 (100.0) 524 (100.0) 525 (100.0) 1049 1231 (100.0) 1225 (100.0)
(100.0)
mITT Population 526 (100.0) 524 (100.0) 525 (100.0) 1049 1229 (99.8) 1224 (99.9)
(100.0)
Completed 24 Weeks 471 (89.5) 479 (91.4) 476 (90.7) 955 (91.0) 1095 (89.0) 1081 (88.2)
Completed 52 Weeks 443 (84.2) 454 (86.6) 448 (85.3) 902 (86.0) NA NA
a. Duration included in the analyses of ES2 pooled population is 24 weeks.

Outcomes and estimation

Primary efficacy endpoint: change from Baseline to Week 24 in A1C

Change from Baseline to Week 24 in A1C was the primary efficacy endpoint in Studies 309 and 310, while
it was the first secondary efficacy endpoint in Study 312. In all studies and all pools, the change from
Baseline in all active groups and the difference versus placebo were statistically significant (p <0.001)
(Table 13).

The change from Baseline to Week 24, and LS mean differences in A1C (%) for ES1 and ES2 (mITT
population) was consistent with individual Phase 3 study results. In the ES1 pool, the LS mean change
from Baseline was ­0.41% and ­0.43% for sotagliflozin 200 mg and 400 mg, the Week 24
placebo-subtracted LS mean A1C reduction was ­0.36% and ­0.38%, respectively, (p <0.001 for both).
Similar results were observed for the ES2 pool at Week 24 (­0.43% for sotagliflozin 400 mg; p <0.001).

Assessment report
EMA/178275/2019 Page 57/125
Table 13 - Analysis of change from Baseline to Week 24 in A1C (%) in the Phase 3 studies
(mITT population)

Study 309 Study 310 Study 312

Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N =262) (N = 258) (N = 261) (N = 263) (N = 703) (N = 699)
Screening
Mean (SD) 8.21 8.26 8.20 8.42 8.35 8.38 8.44 8.48
(0.853) (0.864) (0.847) (0.936) (0.922) (0.937) (0.956) (0.944)
Change from Screening at
Baseline
Mean (SD) -0.66 -0.65 -0.64 -0.64 -0.61 -0.67 -0.23 -0.22
(0.567) (0.538) (0.552) (0.644) (0.666) (0.548) (0.470) (0.553)
Baseline
Mean (SD) 7.54 7.61 7.56 7.79 7.74 7.71 8.21 8.26
(0.712) (0.735) (0.724) (0.881) (0.806) (0.819) (0.921) (0.965)
Week 24
Mean (SD) 7.50 7.17 7.08 7.79 7.36 7.35 7.88 7.41
(0.755) (0.713) (0.725) (0.927) (0.889) (0.906) (1.031) (0.952)
Change from Baseline at
Week 24
Mean (SD) -0.06 -0.42 -0.49 -0.01 -0.39 -0.39 -0.31 -0.81
(0.519) (0.570) (0.566) (0.654) (0.702) (0.654) (0.775) (0.783)
MMRM Model Statistics
LS Mean (SE) -0.07 -0.43 -0.48 -0.02 -0.39 -0.37 -0.33 -0.79
(0.036) (0.036) (0.036) (0.044) (0.044) (0.043) (0.031) (0.032)
95% CLs for change from (-0.14, (-0.50, (-0.56, (-0.11, (-0.47, (-0.46, (-0.39, (-0.85.
Baseline -0.00) -0.36) -0.41) 0.07) -0.30) -0.29) -0.27) -0.73)
p-value 0.038 <0.001 <0.001 0.63 <0.001 <0.001 <0.001 <0.001
Summary of Treatment
Comparison
LS Mean Difference (SE) NA -0.36 -0.41 NA -0.37 -0.35 NA -0.46
from Placebo (0.047) (0.047) (0.058) (0.058) (0.042)
95% CLs for Difference NA (-0.45, (-0.50, NA (-0.48, (-0.47, NA (-0.54,
-0.27) -0.32) -0.25) -0.24) -0.38)
p-value NA <0.001 <0.001 NA <0.001 <0.001 NA <0.001

Secondary efficacy endpoints

Net benefit

Net benefit (defined as the proportion of patients with A1C <7.0% and no episode of SH and no episode
of DKA from Randomization to Week 24) was the primary endpoint in Study 312, and it was the first
secondary efficacy endpoint in Studies 309 and 310.

The proportion of patients with A1C <7.0% at Week 24 is the main component of this composite endpoint.
At Week 24, the difference versus placebo for the responders in net benefit was statistically significant in
all studies and all groups (p <0.001). Table 14 summarizes the results for the Phase 3 studies.

In the ES1 pool, the absolute difference compared to placebo at Week 24 was 14.0% for sotagliflozin 200
mg and 19.5% for sotagliflozin 400 mg (p <0.001 for both). Similar results were observed for the ES2
pool at Week 24 (16.0% difference from placebo for sotagliflozin 400 mg; p <0.001).

Assessment report
EMA/178275/2019 Page 58/125
Table 14 - Analysis of net benefit at Week 24 for the Phase 3 Studies (mITT population)

Study 309 Study 310 Study 312

Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = (N = 258) (N = 261 (N = 263) (N = (N = 699)
262) 703)
Components of Net Benefit
A1C <7.0% at Baseline, n (%) 51 (19.0) 50 (19.0) 51 (19.5) 44 (17.1) 50 (19.2) 46 (17.5) 31 (4.4) 29 (4.1)
A1C <7.0% at Week 24, n (%) 61 (22.8) 97 (36.9) 123 39 (15.1) 87 (33.3) 89 (33.8) 111 207 (29.6)
(46.9) (15.8)
A1C <7.0% at Week 24 and No 58 (21.6) 89 (33.8) 117 39 (15.1) 83 (31.8) 85 (32.3) 107 202 (28.9)
SH, n (%) (44.7) (15.2)
A1C <7.0% at Week 24 and No 61 (22.8) 96 (36.5) 120 39 (15.1) 86 (33.0) 89 (33.8) 111 204 (29.2)
DKA from randomization to (45.8) (15.8)
Week 24, n (%)
Analysis of Net Benefit at Week 24
A1C <7.0% at Week 24 and no
SH and no DKA from
randomization to Week 24
Responders, n (%) 58 (21.6) 88 (33.5) 114 39 (15.1) 82 (31.4) 85 (32.3) 107 200 (28.6)
(43.5) (15.2)
Net benefit difference versus NA 11.8 21.9 NA 16.3 17.2 NA 13.4
placebo
95% CI NA (4.28, (14.10, NA (9.17, (10.06, NA (9.12,
19.36) 29.64) 23.43) 24.35) 17.67)
p-value NA 0.002 <0.001 NA <0.001 <0.001 NA <0.001

Postprandial glucose

Two-hour PPG after a standardized mixed meal was evaluated in patients participating in the CGM
sub-study of Studies 309 and 310 and is primarily evaluated in the ES1 pool (Table 15).
Table 15 - Analysis of change from Baseline to Week 24 in 2-hour postprandial
glucose (mg/dL) in the ES1 Pool (mITT CGM population)

Placebo SOTA SOTA

(N = 93) 200 mg 400 mg
(N = 89) (N = 96)
Baseline
Mean (SD) 229.8 (98.88) 211.5 (99.15) 209.7 (85.69)
Week 24
Mean (SD) 215.1 (86.95) 181.8 (79.11) 171.3 (62.86)
Median (Min, Max) 225.0 (50, 409) 173.0 (58, 479) 170.0 (43, 339)
Change from Baseline at Week 24
Mean (SD) -9.7 (110.26) -39.0 (87.38) -35.0 (84.53)
ANCOVA Model; Statistics
LS Mean (SE) -7.9 (9.38) -42.7 (9.66) -48.8 (9.10)
95% CLs for Difference (-26.4, 10.6) (-61.7, -23.6) (-66.8, -30.9)
p-value 0.40 <0.001 <0.001
Summary of Treatment Comparison
LS Mean Difference (SE) from Placebo NA -34.8 (11.85) -41.0 (11.50)
95% CLs for Difference NA (-58.1, -11.4) (-63.6, -18.3)
p-value NA 0.004 <0.001

Body weight

Assessment report
EMA/178275/2019 Page 59/125
 More than 70% of patients included in the Phase 3 studies were overweight or obese. Table 16
summarizes the change from Baseline to Week 24 in body weight in the Phase 3 studies.

Pooled analyses showed results consistent with the individual Phase 3 studies. In the ES1 pool, the LS
mean change from Baseline at Week 24 was ­1.70 kg and ­2.55 kg for sotagliflozin 200 mg and 400 mg,
respectively. The LS mean difference from placebo in body weight change at Week 24 was ­2.17 for 200
mg and ­3.02 kg for 400 mg respectively (p <0.001 for both). Similar results were observed for the ES2
pool at Week 24 (-2.99 kg LS mean difference from placebo for sotagliflozin 400 mg; p <0.001). In ES1,
18.70% (200 mg) and 23.81% (400 mg) of patients had a weight loss >5% at Week 24 compared to
4.18% with placebo (p <0.001 for both doses).
Table 16 - Analysis of change from Baseline to Week 24 in absolute body weight (kg) in Phase
3 Studies (mITT population)

Study 309 Study 310 Study 312

Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263) (N = 703) (N = 699)
Screening
Mean (SD) 86.39 86.01 85.76 80.46 81.25 81.28 81.07 82.11
(17.554) (18.291) (17.900) (16.685) (17.146) (17.376) (16.901) (17.039)
Change from Screening
at Baseline
Mean (SD) 0.90 0.95 0.74 0.62 0.69 0.68 (2.072) 0.48 0.30
(1.894) (2.190) (1.904) (1.925) (1.867) (1.546) (1.530)
Baseline
Mean (SD) 87.30 86.96 86.50 81.08 81.93 81.97 81.55 82.40
(17.709) (18.539) (18.004) (16.857) (17.386) (17.963) (17.032) (17.131)
Week 24
Mean (SD) 87.91 85.00 84.15 81.48 80.05 80.09 82.70 80.19
(17.665) (18.494) (17.680) (16.978) (17.384) (17.919) (17.513) (17.017)
Change from Baseline at
Week 24
Mean (SD) 0.78 -1.68 -2.72 0.17 -1.88 -2.40 0.71 -2.33
(2.745) (2.928) (2.863) (2.684) (2.990) (3.373) (2.924) (3.044)
MMRM Model Statistics
LS mean (SE) 0.78 -1.57 -2.67 0.11 -1.88 -2.47 0.77 -2.21
(0.187) (0.188) (0.188) (0.201) (0.200)) (0.199) (0.122) (0.122)
95% CLs for change (0.41, 1.15) (-1.94, (-3.04, (-0.29, (-2.27, (-2.86, (0.53, (-2.45,
from Baseline -1.20) -2.30) 0.50) -1.49) -2.08) 1.01) -1.97)
p-value <0.001 <0.001 <0.001 0.60 <0.001 <0.001 <0.001 <0.001
Summary of treatment
comparison
LS mean difference NA -2.35 -3.45 NA -1.98 -2.58 NA -2.98
(SE) from Placebo (0.256) (0.256) (0.276) (0.276) (0.166)
95% CLs for NA (-2.85, (-3.95, NA (-2.53, (-3.12, NA (-3.31,
difference -1.85) -2.94) -1.44) -2.04) -2.66)
p-value NA <0.001 <0.001 NA <0.001 <0.001 NA <0.001

Mean daily insulin dose

In Studies 309 and 310, the mean total daily insulin dose at Screening ranged between 58.7 and 64.0
IU/day. During the 6-week Run-in Period with insulin optimization, the mean Baseline total daily insulin
dose increased up to 60.3 to 66.8 IU/day, with individual doses ranging between 11.9 and 290.0 IU/day.

In Study 312 (conducted without insulin optimization), the mean total daily insulin dose at Baseline was
approximately 57 IU/day, with individual doses ranging between 6.5 and 253.0 IU/day.

Assessment report
EMA/178275/2019 Page 60/125
Because of the high range of insulin dose expressed as IU/day at Baseline, the changes from Baseline in
insulin doses are mainly expressed as percent change from Baseline. While change in bolus insulin dose
at Week 24 was the secondary endpoint, results for basal and total insulin dose are also presented here
for completeness.

Daily bolus insulin dose

In Studies 309 and 310, the mean Baseline bolus insulin dose after the 6-week Run-in Period with insulin
optimization ranged between 30.3 and 32.1 IU/day. In the placebo groups, a further LS mean increase
from Baseline of +3.9% and +5.9%, respectively was observed at Week 24. In the sotagliflozin groups,
a LS mean decrease was observed, ranging between ­1.8% and ­10.5%, in parallel of an improvement in
glucose control. In Studies 309 and 310, the LS mean difference from placebo was ­5.7% (p = 0.12) and
­12.9% (p <0.001) with sotagliflozin 200 mg; it was ­12.7% and ­16.4% (both p <0.001) with
sotagliflozin 400 mg (Figure 14 and Figure 15).

Similar results were observed in Study 312, with an LS mean increase of +6.6% in the placebo group and
a LS mean decrease of ­5.7% in the sotagliflozin 400 mg group. The LS mean difference from placebo at
Week 24 was ­12.3% (p <0.001) (Figure 16).

Figure 14, Figure 15, and Figure 16 below provide mean percent change in Total, Basal, and Bolus insulin
doses at Week 24 observed in Studies 309, 310, and 312, respectively.
Figure 14 - LS mean percent change in total, basal and bolus insulin doses at Week 24 in
Study 309 (mITT population)

Assessment report
EMA/178275/2019 Page 61/125
Figure 15 - LS mean percent change in total, basal and bolus insulin doses at Week 24 in
Study 310 (mITT population)

Figure 16 - LS mean percent change in total, basal and bolus insulin doses at Week 24 in
Study 312 (mITT population)

The analysis of percent change from Baseline to Week 24 in mean daily bolus insulin dose in the ES1 and
ES2 pools (mITT population) showed results consistent with the individual Phase 3 studies. In the ES1
pool, the Week 24 LS mean difference from placebo was -9.29% and -14.39% for sotagliflozin 200 mg
and 400 mg (p <0.001 for both). Similar results were observed for the ES2 pool at Week 24 (-13.27% LS
mean difference from placebo for sotagliflozin 400 mg; p <0.001).

Daily basal insulin dose

In Studies 309 and 310, the mean Baseline basal insulin dose after insulin optimization Run-in ranged
between 29.2 and 35.1 IU/day. In the placebo groups, a further LS mean increase at Week 24 of +3.8%

Assessment report
EMA/178275/2019 Page 62/125
and +1.7% from Baseline was observed, showing that insulin optimization was maintained during the
study. In the sotagliflozin groups, a slight LS mean decrease was observed, ranging between ­1.7% and
­5.4%, in parallel with an improvement in glucose control. The LS mean difference from placebo with
sotagliflozin ranged between ­4.7% (p = 0.030) and ­9.1% (p <0.001) (Figure 14 and Figure 15).

Similar results were observed in Study 312 with a LS mean increase of +6.8% in the placebo group and
a decrease of ­3.1% in the sotagliflozin 400 mg group. The LS mean difference from placebo at Week 24
was ­9.9%; p <0.001 (Figure 16).

Consistent results were observed in the ES1 pool (Week 24 placebo-subtracted LS mean reduction of
­5.7% and of ­6.9% for the sotagliflozin 200 mg and 400 mg groups; p <0.001 for both) and for the ES2
pool at Week 24 (­8.6% LS mean difference from placebo for sotagliflozin 400 mg; p <0.001).

Total daily insulin dose

Overall, a slight but statistically significant decrease in the mean total insulin dose was observed,
consistent across the Phase 3 program. In all 3 Phase 3 studies the pattern of total insulin dose reductions
was just under 10% compared to placebo, with the percent reductions being somewhat greater for bolus
than basal insulin (Figure 14, Figure 15 and Figure 16).

Consistent results were observed in the ES1 pool (Week 24 placebo-subtracted LS mean reduction ­7.2%
and ­9.6% for the sotagliflozin 200 mg and 400 mg groups; p <0.001 for both) and for the ES2 pool at
Week 24 (­9.7% LS mean difference from placebo for sotagliflozin 400 mg; p <0.001).

Fasting plasma glucose

At Baseline, the mean FPG was between 148.2 and 165.5 mg/dL in the 3 Phase 3 studies. At Week 24, the
LS mean difference from placebo in FPG was statistically significant for all sotagliflozin groups, ranging
from ­9.8 to ­25.7 mg/dL (p <0.001 for all groups with the exception of sotagliflozin 200 mg in Study 309
[p = 0.036]).

Results for the ES1 and ES2 pools were consistent with the Phase 3 study results. In the ES1 pool at Week
24, the LS mean difference from placebo was ­15.7 mg/dL and ­21.4 mg/dL for the sotagliflozin 200 mg
and 400 mg (p <0.001 for both) and for the ES2 pool at Week 24, the LS mean difference from placebo
was ­22.3 mg/dL with sotagliflozin 400 mg (p <0.001).

Patient reported outcomes

Diabetes treatment satisfaction as measured by DTSQs

The change from Baseline to Week 24 in DTSQs scores was analysed in Studies 309 and 310. The DTSQs
total scores range from 0 to 36, with higher scores indicating higher satisfaction. The LS mean difference
compared to placebo ranged between +1.7 and +2.5 in all sotagliflozin groups (all p <0.001), indicating
an improved satisfaction with sotagliflozin.

The DTSQs responder rate (defined as a change in score of 3 or more as threshold) for the 200 mg and
400 mg sotagliflozin groups were significantly higher (p <0.005) for both groups compared to placebo in
both Studies 309 and 310. In Study 309, the responder rates for the 200 mg and 400 mg doses were
53.7% and 52.0%, respectively, compared with 27.9% for placebo. In Study 310, the responder rates for
the 200 mg and 400 mg doses were 40.9% and 43.0%, respectively, compared with 28.8% in the
placebo.

Diabetes distress as measured by 2-item DDS2 scores

The change from Baseline to Week 24 in DDS2 scores was analysed in Studies 309 and 310. The DDS2 is
a validated 2-item diabetes distress screening instrument requesting respondents to rate, on a 6-point
scale, the degree of their distress. A higher score indicates higher distress. At Week 24, the LS mean

Assessment report
EMA/178275/2019 Page 63/125
difference compared to placebo ranged between ­0.3 (p = 0.025) and ­0.8 (p <0.001) in all sotagliflozin
groups, indicating a lower distress with sotagliflozin.

Other efficacy endpoints

Systolic blood pressure

Changes in SBP were evaluated in Studies 309 and 310 at Week 12, consistent with protocol
recommendation to keep doses of antihypertensive medications stable until Week 12 to allow for
appropriate blood pressure assessment. For Study 312, the SBP evaluation was performed at Week 16 for
similar reasons.

Modest reductions in SBP were observed at Week 12 for the overall population for both sotagliflozin doses
compared to placebo in all Phase 3 studies as noted in Table 17. Reductions in SBP were maintained at
Week 24 for both sotagliflozin doses in all Phase 3 studies.

Consistent results were observed for change in SBP in the overall population in the pooled ES1 population
at Week 12, with a placebo-adjusted change in SBP of -2 mmHg (95%CI -3.2, -0.7) in the 200 mg group
and -3.5 mmHg (95%CI -4.7, -2.3) in the 400 mg group. Reductions in SBP were maintained at Week 24
for both sotagliflozin doses.
Table 17 - Change from Baseline in systolic blood pressure (mm Hg) at Week 12 in the Studies
309 and 310 and at Week 16 in Study 312 in the overall population (mITT population)

Study 309 Study 310 Study 312

Placebo SOTA SOTA Placebo SOTA SOTA Placebo SOTA
200 mg 400 mg 200 mg 400 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263) (N = 703) (N = 699)
Baseline
Mean (SD) 120.9 120.0 119.5 123.1 123.0 123.1 121.8 122.0
(13.47) (14.84) (14.73) (15.53) (15.08) (13.69) (14.82) (15.25)
Week 12 (309 and 310)
Week 16 (312)
Mean (SD) 121.2 117.2 116.6 121.6 121.4 118.3 122.5 118.7
(13.25) (13.89) (12.99) (14.74) (14.50) (13.54) (14.83) (14.47)
Change from Baseline
at Week 12 for 309 and
310; at Week 16 for 312
Mean (SD) 0.6 (10.89) -2.6 -3.3 -1.8 -2.2 -4.8 (10.81) 0.5 (11.36) -3.3
(12.20) (10.27) (10.07) (11.59) (11.31)
MMRM Model;
Statistics
LS Mean (SE) 1.0 (0.66) -2.5 (0.67) -3.2 (0.66) -2.4 (0.68) -2.8 (0.67) -5.2 (0.67) 0.3 (0.43) -3.5 (0.44)
95% CLs for change (-0.3, 2.3) (-3.8, -1.2) (-4.5, -1.9) (-3.7, -1.1) (-4.1, -1.5) (-6.5, -3.9) (-0.6, 1.1) (-4.4, -2.7)
from Baseline
p-value 0.14 <0.001 <0.001 <0.001 <0.001 <0.001 0.53 <0.001
Summary of Treatment
Comparisons
LS Mean Difference NA -3.5 (0.88) -4.2 (0.88) NA -0.4 (0.89) -2.8 (0.89) NA -3.8 (0.57)
from Placebo
95% CLs for NA (-5.2, -1.8) (-5.9, -2.4) NA (-2.2, 1.3) (-4.6, -1.1) NA (-4.9, -2.7)
difference
p-value NA <0.001 <0.001 NA 0.64 0.001 NA <0.001

Assessment report
EMA/178275/2019 Page 64/125
SBP results in subset of patients with Baseline SBP ≥130 mm Hg:

Change from Baseline at Week 16 in SBP in the subset of patients with Baseline SBP ≥130 mm Hg was a
secondary efficacy endpoint in Study 312. The LS mean difference from placebo at Week 16 was ­3.5 mm
Hg (p = 0.002). The statistically significant effect was maintained at Week 24 (LS mean difference
­3.6 mm Hg; (p = 0.003). SBP was also evaluated in the subset of patients with Baseline SBP ≥140 mm
Hg. There were 84 patients in the placebo group and 98 patients in the sotagliflozin 400 mg group for
evaluation in this subset. Statistically significant reductions in SBP at Week 16 of ­6.5 mm Hg (p <0.001)
were observed for this subset of patients.

In studies 309 and 310, Change from Baseline in SBP in the subset of patients with Baseline SBP
≥130 mm Hg was also evaluated at Week 12 and at Week 24. To increase precision of the analysis in this
subpopulation, results are presented for the ES1 pool. The LS mean difference from placebo in SBP at
Week 12 for sotagliflozin 200 mg and sotagliflozin 400 mg was ­3.6 mm Hg (p = 0.011) and ­5.3 mm Hg
(p <0.001). The LS mean difference from placebo was maintained at Week 24 with SBP reduction of
­3.1 mm Hg (p = 0.031) for sotagliflozin 200 mg and ­3.2 mm Hg (p = 0.026) for sotagliflozin 400 mg.

CGM sub-study

Pooled data from studies 309 and 310 was prespecified as the main evaluable dataset for the CGM
substudy. A total of 278 patients were randomized, n = 93 placebo, n = 89 sotagliflozin 200 mg, n = 96
sotagliflozin 400 mg. Of the patients included, 49.2% were using CSII as their insulin delivery method.

Primary CGM efficacy endpoint: percent time spent inside target range (70-180 mg/dL)

The percent time spent inside the target range for CGM glucose (70-180 mg/dL) increased from 52.2% to
57.8% with sotagliflozin 200 mg and from 50.7% to 64.2% with sotagliflozin 400 mg from Baseline to
Week 24, whereas no relevant change was observed for placebo. In both sotagliflozin dose groups, the
increase in percent time in range was associated with a significant decrease in the percent time spent
above 180 mg/dL, while the percent time spent below 70 mg/dL was not increased (Figure 17, Table 18).
Assuming 100% daily CGM data is available for analysis, 1.0% of daily CGM time corresponds to
approximately 15 minutes (0.24 hours).

Assessment report
EMA/178275/2019 Page 65/125
Figure 17 - Percent time spent in range and outside range for CGM glucose in the pooled 309
and 310 studies (mITT CGM population)

p-values are for LS mean differences from placebo.

Secondary and other CGM endpoints

Secondary CGM endpoints confirmed that sotagliflozin 200 and 400 mg qd decreased the time spent in
hyperglycaemia. Sotagliflozin therapy resulted in increased time in range with no increase in
hypoglycaemia. The time spent in hypoglycaemia (<55 mg/dL or <70 mg/dL) was similar among
treatment groups, although it was numerically lower with sotagliflozin.

Significant differences were also observed for various measures of glycaemic variability including mean
daily glucose, SD and MAGE (Table 18).
Table 18 - CGM primary, key secondary and other endpoints (mITT CGM population)

Pooled 309 and 310

Placebo SOTA SOTA
200 mg 400 mg
(N = 93) (N = 89) (N = 96)
Percent Time of CGM Glucose 70-180 mg/dL
Mean Baseline (SD) 52.301 (13.7950) 52.193 (15.2656) 50.656 (14.7623)
Mean Week 24 (SD) 51.577 (14.7206) 57.779 (15.8932) 64.172 (13.9760)
LS mean change from Baseline at Week 24 (SE) -1.261 (1.8140) 4.088 (1.8017) 10.448 (1.6964)
95% CLs for change from Baseline (-4.836, 2.314) (0.537, 7.638) (7.104, 13.791)
p-value 0.49 0.024 <0.001
LS mean difference (SE) from Placebo NA 5.349 (2.3883) 11.709 (2.3159)
95% CLs for difference NA (0.639, 10.059) (7.141, 16.276)
p-value NA 0.026 <0.001

Assessment report
EMA/178275/2019 Page 66/125
 Pooled 309 and 310
Placebo SOTA SOTA
200 mg 400 mg
(N = 93) (N = 89) (N = 96)
Percent Time of CGM Glucose <55 mg/dL
LS mean difference (SE) from Placebo NA -0.161 (0.4953) -0.441 (0.4801)
95% CLs for difference NA (-1.137, 0.816) (-1.388, 0.505)
p-value NA 0.75 0.36
Percent Time of CGM Glucose <70 mg/dL
LS mean difference (SE) from Placebo NA -0.330 (0.8618) -0.068 (0.8375)
95% CLs for difference NA (-2.030, 1.369) (-1.720, 1.583)
p-value NA 0.70 0.93
Percent Time of CGM Glucose >180 mg/dL
LS mean difference (SE) from Placebo NA -5.036 (2.6101) -11.746 (2.5332)
95% CLs for difference NA (-10.183, 0.112) (-16.742, -6.750)
p-value NA 0.055 <0.001
Percent Time of CGM Glucose >250 mg/dL
LS mean difference (SE) from Placebo NA -3.662 (1.8179) -7.695 (1.7624)
95% CLs for difference NA (-7.248, -0.077) (-11.171, -4.219)
p-value NA 0.045 <0.001
Mean Daily Glucose (mg/dL)
LS mean difference (SE) from Placebo NA -7.939 (4.7010) -18.928 (4.5586)
95% CLs for difference NA (-17.210, 1.332) (-27.918, -9.938)
p-value NA 0.09 <0.001
MAGE
LS mean difference (SE) from Placebo NA -12.702 (5.5164) -22.111 (5.3718)
95% CLs for difference NA (-23.581, -1.822) (-32.705, -11.517)
p-value NA 0.022 <0.001
CGM Standard Deviation
LS mean difference (SE) from Placebo NA -4.591 (2.2445) -6.781 (2.1881)
95% CLs for difference NA (-9.017, -0.164) (-11.097, -2.466)
p-value NA 0.042 0.002
The CGM tracings of mean 24-hour glucose excursions

The CGM tracings of mean 24-hour glucose excursions (data collected from midnight [00:00]), consisting
of interstitial glucose readings collected every 5 minutes from the week prior to the Week 24 visit are
shown in Figure 18 below. Each line represents mean values from each treatment group (triangles =
placebo [n=93]; squares = sotagliflozin 200 mg [n=89]; diamonds = sotagliflozin 400 mg [n=96]). Top
of target CGM range = 10.0 mmol/L (180 mg/dL).

Compared with placebo the glycaemic curve is pulled down in both preprandial and postprandial periods.
The decline in mean glucose value from 12 AM to 6 AM in sotagliflozin-treated patients was not associated
with a significant increase in nocturnal hypoglycaemia.

Figure 18 - CGM tracings of mean 24-hour glucose excursions

Assessment report
EMA/178275/2019 Page 67/125
DEXA sub-study

Pooled data from studies 309 and 310 were prespecified as the main evaluable dataset. The mITT
population for the ES1 DEXA sub-study included 75 placebo patients, 84 sotagliflozin 200 mg patients,
and 84 sotagliflozin 400 mg patients recruited from selected sites participating in the DEXA sub-study.
Pooled fat mass and lean body mass data from the 309 and 310 DEXA substudies are provided below as
the primary dataset for DEXA endpoints.

The primary objective for the pooled analysis was to compare the effect of sotagliflozin versus placebo on
total fat mass at Week 24, and a decrease was noted for both sotagliflozin groups compared to an
increase with placebo. The LS mean difference from placebo at Week 24 was ­2188.12 g, p <0.001 for
sotagliflozin 200 mg and ­2515.95 g, p <0.001 for sotagliflozin 400 mg group. Similar results were
observed expressing the data as percent total fat mass (LS mean difference ­1.34%, p <0.001 for
sotagliflozin 200 mg and ­1.56%, p <0.001 for sotagliflozin 400 mg).

A modest but statistically significant decrease in total lean mass from Baseline to Week 24 was also noted
for both sotagliflozin groups. The LS mean difference from placebo at Week 24 was ­771.01 g, p = 0.015
for sotagliflozin 200 mg and ­913.33 g, p = 0.003 for sotagliflozin 400 mg.

Ancillary analyses

Long-Term Effects on A1C

Figure 19 and Figure 20 show the kinetics of A1C during the course of Studies 309 and 310.
Approximately 80% of the effect was observed during the first 4 weeks and the effect was maximum at
Week 8 and Week 12. Although A1C then slightly increased in all groups, there was still a difference
versus placebo at Week 52.

Assessment report
EMA/178275/2019 Page 68/125
Figure 19 - Change in A1C from Baseline to Week 52 in Study 309 (mITT population)

Figure 20 - Change in A1C from Baseline to Week 52 in Study 310 (mITT population)

The LS mean difference from placebo remained statistically significant in all groups at Week 52, ranging
between ­0.21 to ­0.32%. For the ES1 pool (mITT population), the LS mean difference from placebo was
­0.23% for sotagliflozin 200 mg and ­0.32% for sotagliflozin 400 mg (both p <0.001).

Long-Term Effects on Other Efficacy Parameters

Net benefit

The difference from placebo for the net benefit remained statistically significant in all groups in the Phase
3 Studies 309 and 310 at Week 52, ranging between 7.2 to 13.4% (Table 19). For the ES1 pool (mITT

Assessment report
EMA/178275/2019 Page 69/125
population), the LS mean difference from placebo was 9.2% for sotagliflozin 200 mg and 12.8% for
sotagliflozin 400 mg (both p <0.001).
Table 19 - Analysis of the components of net benefit and analysis of net benefit at Week 52
for Study 309 and Study 310 (mITT population)

Study 309 Study 310

Placebo SOTA SOTA Placebo SOTA SOTA
200 mg 400 mg 200 mg 400 mg
(N = 268) (N = 263) (N = 262) (N = 258) (N = 261) (N = 263)
Components of Net Benefit
Baseline
A1C <7.0% at 51 (19.0) 50 (19.0) 51 (19.5) 44 (17.1) 50 (19.2) 46 (17.5)
Baseline, n (%)
Week 52
A1C <7.0% at 56 (20.9) 79 (30.0) 93 (35.5) 40 (15.5) 71 (27.2) 73 (27.8)
Week 52, n (%)
A1C <7.0% at Week 51 (19.0) 72 (27.4) 87 (33.2) 37 (14.3) 67 (25.7) 70 (26.6)
52 and No SH from
Baseline to Week 52,
n (%)
A1C <7.0% at Week 55 (20.5) 75 (28.5) 91 (34.7) 40 (15.5) 71 (27.2) 72 (27.4)
52 and No DKA from
Baseline to Week 52,
n (%)
Analysis of Net Benefit at Week 52
Patients with A1C <7.0% and no SH and no DKA
Responders, n (%) 51 (19.0) 69 (26.2) 85 (32.4) 37 (14.3) 67 (25.7) 70 (26.6)
Treatment Comparison
Net benefit NA 7.2 13.4 NA 11.3 12.3
difference in %
responders versus
placebo
95% CI NA (0.11, 14.30) (6.05, 20.78) NA (4.52, 18.14) (5.43, 19.12)
p-value NA 0.049 <0.001 NA 0.001 <0.001

The subgroup analyses by demographics and Baseline factors at Week 52 confirmed that the LS mean
difference from placebo at Week 52 was statistically significant at both doses in all subgroups of sufficient
size.

Body weight

Figure 21 shows the kinetics of body weight reduction during the course of Studies 309 and 310. Weight
reduction was progressive and sustained throughout the study. In Study 309, the Baseline mean body
weight was 87.0 kg and 86.5 kg for patients on sotagliflozin 200 mg and 400 mg, respectively. Baseline
mean body weight in Study 310 was 81.93 kg and 81.97 kg for sotagliflozin 200 mg and 400 mg,
respectively. After Week 24, body weight continued to decrease or stabilized up to the end of the study.

The LS mean difference from placebo remained statistically significant in all groups at Week 52, ranging
between ­2.18 to ­4.32 kg.

Assessment report
EMA/178275/2019 Page 70/125
Figure 21 - Change from Baseline to Week 52 in absolute body weight (kg) in Study 309 (top
with Baseline) and Study 310 (bottom) (mITT population)

For the ES1 pool (mITT population), the LS mean difference from placebo was ­2.67 kg for sotagliflozin
200 mg and ­3.64 kg for sotagliflozin 400 mg (both p <0.001). This corresponded to a -3.21% (200 mg)
and -4.24% LS mean percent change from Baseline to Week 24 (both p <0.001).

Mean daily insulin dose

The LS mean difference from placebo in percent change in bolus insulin dose remained statistically
significant in all groups in Study 309 and Study 310 at Week 52, ranging between -5.53 to -15.63%.

For the ES1 pool (mITT population), the LS mean difference from placebo in percent change in bolus
insulin dose was -6.63% for sotagliflozin 200 mg (p = 0.037) and -13.73% for sotagliflozin 400 mg
(p <0.001).

Similar results were observed for the mean daily basal insulin dose and for the mean daily total insulin
dose, see Figure 22 and Figure 23.

Assessment report
EMA/178275/2019 Page 71/125
Figure 22 - Percent change from Baseline to Week 52 in mean total daily insulin dose, basal
insulin dose and bolus insulin dose (IU/day) in Study 309 (mITT population)

Figure 23 - Percent change from Baseline to Week 52 in mean total daily insulin dose, basal
insulin dose, and bolus insulin dose (IU/day) in Study 310 (mITT population)

Fasting plasma glucose

Consistent with change from Baseline to Week 24, the LS mean difference from placebo remained
statistically significant in most groups at Week 52 in Study 309 and Study 310, ranging between ­4.9 to

Assessment report
EMA/178275/2019 Page 72/125
­19.4 mg/dL. For the ES1 pool (mITT population), the LS mean difference from placebo was ­8.7 mg/dL
for sotagliflozin 200 mg (p = 0.032) and ­17.7 mg/dL for sotagliflozin 400 mg (p <0.001).

Diabetes distress as measured by 2-item DDS2 scores

At Week 52, the LS mean difference in DDS2 scores compared to placebo remained statistically significant
for sotagliflozin 400 mg in Study 310 and for the 2 doses in Study 309. For the ES1 pool (mITT
population), the LS mean difference from placebo was statistically significant for sotagliflozin 200 mg (p
= 0.003) and sotagliflozin 400 mg (p <0.001), which confirmed the persistence of the beneficial effect on
patient distress.

Systolic blood pressure

Consistent with change from Baseline to Week 24, there was a modest but statistically significant
decrease in the change from Baseline to Week 52 in SBP (mm Hg) in both sotagliflozin groups compared
with placebo in Studies 309 and 310 and in the ES1 pool. In the ES1 pool, the LS mean difference from
placebo at Week 52 was ­2.9 mm Hg (p <0.001) for sotagliflozin 200 mg and -3.6 mm Hg (p <0.001) for
sotagliflozin 400 mg.

DEXA substudy

The analysis of change from Baseline to Week 52 in DEXA total fat mass in the ES1 pool (mITT DEXA
Population) demonstrated the continued persistence of efficacy of sotagliflozin. The LS mean difference
(SE) from placebo at Week 52 was ­1695.72 g (509.962), p = 0.001 for sotagliflozin 200 mg and
­2123.19 g (499.843), p <0.001 for sotagliflozin 400 mg, confirming that the sustained beneficial effect
on body weight was primarily due to a reduction in fat mass.

Summary of main efficacy results

The following tables summarise the efficacy results from the main studies supporting the present
application. These summaries should be read in conjunction with the discussion on clinical efficacy as well
as the benefit risk assessment (see later sections).
Table 20 Summary of efficacy for trial 309

Title: A phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study to evaluate the safety, efficacy,
and tolerability of LX4211 as adjunct therapy in adult patients with type 1 diabetes mellitus who have inadequate glycemic control
with insulin therapy
Study identifier LX4211.309
Design United States and Canada, randomized, double-blind, 3-arm, equal allocation, parallel-group
Duration of main treatment phase 24 weeks
(Double-blind Core Treatment):
Duration of run-in phase: 6 weeks
Duration of long-term extension 24 weeks
(Double-blind LTE):
Total duration of treatment: 52 weeks
Hypothesis Superiority
Treatment groups Placebo once daily 268 patients randomized
793 patients randomized Sotagliflozin 200 mg once daily 263 patients randomized
Sotagliflozin 400 mg once daily 262 patients randomized
Endpoints and definitions Primary A1C (%) Change from Baseline in A1C to Week 24
endpoint

Assessment report
EMA/178275/2019 Page 73/125
Title: A phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study to evaluate the safety, efficacy,
and tolerability of LX4211 as adjunct therapy in adult patients with type 1 diabetes mellitus who have inadequate glycemic control
with insulin therapy
Study identifier LX4211.309
Key Net benefit Proportion of patients with A1C <7.0% at Week 24 and no
secondary episode of severe hypoglycaemia or DKA during the main
endpoints treatment phase

Body weight (kg) Change from Baseline in body weight at Week 24

Bolus insulin dose Change from Baseline in mean daily bolus insulin dose at
(IU/day) Week 24

FPG (mg/dL) Change from Baseline in FPG at Week 24

Database lock 15 April 2017
Results and Analysis
Analysis description Primary and Key Secondary Analyses
Analysis population; time mITT population (all randomized patients who took at least 1 dose of study drug); 24 weeks from
point description Baseline (end of main treatment phase)
Descriptive statistics and Treatment group Placebo Sotagliflozin 200 mg Sotagliflozin 400 mg
estimate variability Number of patients 268 263 262
A1C (%) change from -0.07 -0.43 -0.48
Baseline: LSM (SE) (0.036) (0.036) (0.036)
p = 0.038 p <0.001 p <0.001
Net benefit (%) 21.6 33.5 43.5
Body weight (kg) 0.78 -1.57 -2.67
change from Baseline: (0.187) (0.188) (0.188)
LSM (SE) p <0.001 p <0.001 p <0.001
Bolus insulin dose -0.84 -2.33 -4.13
(IU/day) change from (0.688) (0.692) (0.692)
Baseline: LSM (SE) p = 0.22 p <0.001 p <0.001
FPG (mg/dL) change 3.7 -6.1 -14.0
from Baseline: LSM (3.45) (3.47) (3.48)
(SE) p = 0.28 p = 0.08 p <0.001
Effect estimate per Comparison groups Sotagliflozin 200 mg Sotagliflozin 400 mg
comparison versus placebo versus placebo
A1C (%) (primary LSM difference (SE) -0.36 (0.047) -0.41 (0.047)
endpoint)
95% CI -0.45, -0.27 -0.50, -0.32
P-value <0.001 <0.001
Net benefit (%) Difference 11.8 21.9
95% CI 4.28, 19.36 14.10, 29.64
P-value 0.002 <0.001
Body weight (kg) LSM difference (SE) -2.35 (0.256) -3.45 (0.256)
95% CI -2.85, -1.85 -3.95, -2.94
P-value <0.001 <0.001
Bolus insulin dose LSM difference (SE) -1.50 (0.917) -3.30 (0.916)
(IU/day)
95% CI -3.30, 0.30 -5.09, -1.50

Assessment report
EMA/178275/2019 Page 74/125
Title: A phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study to evaluate the safety, efficacy,
and tolerability of LX4211 as adjunct therapy in adult patients with type 1 diabetes mellitus who have inadequate glycemic control
with insulin therapy
Study identifier LX4211.309
P-value 0.10 <0.001
FPG (mg/dL) LSM difference (SE) -9.8 (4.67) -17.8 (4.69)
95% CI -19.0, -0.7 -27.0, -8.6
P-value 0.036 <0.001
Notes The N for individual endpoints and time points may be less than the number randomized to the group.
A1C = glycosylated hemoglobin (HbA1c); CI = confidence interval; FPG = fasting plasma glucose;
IU = international unit; LSM = least squares mean; mITT = modified intent-to-treat; SE = standard error

Table 21 Summary of efficacy for trial 310

Title: A phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study to evaluate the efficacy, safety,
and tolerability of LX4211 as adjunct therapy in adult patients with type 1 diabetes mellitus who have inadequate glycemic control
with insulin therapy
Study identifier LX4211.310
Design Multinational, randomized, double-blind, 3-arm, equal allocation, parallel-group
Duration of main treatment phase 24 weeks
(Double-blind Core Treatment):
Duration of run-in phase: 6 weeks
Duration of long-term extension 24 weeks
(Double-blind LTE):
Total duration of treatment: 52 weeks
Hypothesis Superiority
Treatment groups Placebo once daily 258 patients randomized
782 patients randomized Sotagliflozin 200 mg once daily 261 patients randomized
Sotagliflozin 400 mg once daily 263 patients randomized
Endpoints and definitions Primary A1C (%) Change from Baseline in A1C to Week 24
endpoint
Key Net benefit Proportion of patients with A1C <7.0% at Week 24 and no
secondary episode of severe hypoglycaemia or DKA during the main
endpoints treatment phase

Body weight (kg) Change from Baseline in body weight at Week 24

Bolus insulin dose Change from Baseline in mean daily bolus insulin dose at
(IU/day) Week 24

FPG (mg/dL) Change from Baseline in FPG at Week 24

Database lock 17 July 2017
Results and Analysis
Analysis description Primary and Key Secondary Analyses
Analysis population and mITT population (all randomized patients who took at least 1 dose of study drug); 24 weeks from
time point description Baseline (end of main treatment phase)

Assessment report
EMA/178275/2019 Page 75/125
Title: A phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study to evaluate the efficacy, safety,
and tolerability of LX4211 as adjunct therapy in adult patients with type 1 diabetes mellitus who have inadequate glycemic control
with insulin therapy
Study identifier LX4211.310
Descriptive statistics and Treatment group Placebo Sotagliflozin 200 mg Sotagliflozin 400 mg
estimate variability Number of patients 258 261 263
A1C (%) change from -0.02 -0.39 -0.37
Baseline: LSM (SE) (0.044) (0.044) (0.043)
p = 0.63 p <0.001 p <0.001
Net benefit (%) 15.1 31.4 32.3
Body weight (kg) 0.11 -1.88 -2.47
change from Baseline: (0.201) (0.200) (0.199)
LSM (SE) p = 0.60 p <0.001 p <0.001
Bolus insulin dose -1.19 -4.38 -4.78
(IU/day) change from (0.635) (0.636) (0.634)
Baseline: LSM (SE) p = 0.06 p <0.001 p <0.001
FPG (mg/dL) change 8.8 -12.8 -16.9
from Baseline: LSM (3.95) (3.97) (3.96)
(SE) p = 0.026 p = 0.001 p <0.001
Effect estimate per Comparison groups Sotagliflozin 200 mg Sotagliflozin 400 mg
comparison versus placebo versus placebo
A1C (%) (primary LSM difference (SE) -0.37 (0.058) -0.35 (0.058)
endpoint)
95% CI -0.48, -0.25 -0.47, -0.24
P-value <0.001 <0.001
Net benefit (%) Difference 16.3 17.2
95% CI 9.17, 23.43 10.06, 24.35
P-value <0.001 <0.001
Body weight (kg) LSM difference -1.98 (0.276) -2.58 (0.276)
(SE)
95% CI -2.53, -1.44 -3.12, -2.04
P-value <0.001 <0.001
Bolus insulin dose LSM difference (SE) -3.20 (0.847) -3.59 (0.845)
(IU/day)
95% CI -4.86, -1.53 -5.25, -1.93
P-value <0.001 <0.001
FPG (mg/dL) LSM difference (SE) -21.6 (5.38) -25.7 (5.37)
95% CI -32.2, -11.0 -36.2, -15.1
P-value <0.001 <0.001
Notes The N for individual endpoints and time points may be less than the number randomized to the group.
A1C = glycosylated hemoglobin (HbA1c); CI = confidence interval; FPG = fasting plasma glucose;
IU = international unit; LSM = least squares mean; mITT = modified intent-to-treat; SE = standard error

Assessment report
EMA/178275/2019 Page 76/125
Table 22 Summary of efficacy for trial 312

Title: A phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study to evaluate the net clinical benefit
of sotagliflozin as adjunct to insulin therapy in type 1 diabetes
Study identifier LX4211.312
Design Multinational, randomized, double-blind, 2-arm, equal allocation, parallel-group
Duration of main treatment phase 24 weeks
(Double-blind):
Duration of run-in phase: 2 weeks
Total duration of treatment: 24 weeks
Hypothesis Superiority
Treatments groups Placebo once daily 705 patients randomized
1405 patients randomized Sotagliflozin 400 mg once daily 700 patients randomized
Endpoints and definitions Primary Net benefit Proportion of patients with A1C <7.0% at Week 24 and no
endpoint episode of severe hypoglycaemia or DKA during the main
treatment phase
Key A1C (%) Change from Baseline in A1C at Week 24
secondary
endpoints Body weight (kg) Change from Baseline in body weight at Week 24

SBP (mm Hg) Change from Baseline in SBP at Week 16 in subset of

patients with Baseline SBP ≥130 mm Hg

Bolus insulin dose (IU) Change from Baseline in in mean daily bolus insulin dose at
Week 24
Database lock 12 May 2017
Results and Analysis
Analysis description Primary and Key Secondary Analyses
Analysis population and mITT population (all randomized patients who took at least 1 dose of study drug); 24 weeks from
time point description Baseline (end of main treatment phase)
Descriptive statistics and Treatment group Placebo Sotagliflozin 400 mg
estimate variability Number of patients 703 699
Net benefit (%) 15.2 28.6
A1C (%) change from -0.33 -0.79
Baseline: LSM difference (0.031) (0.032)
(SE) p <0.001 p <0.001
Body weight (kg) change 0.77 -2.21
from Baseline: LSM (0.122) (0.122)
difference (SE) p <0.001 p <0.001
SBP (mm Hg) change Subset N=203 Subset N=203
from Baseline: LSM -5.7 (0.90) -9.2 (0.92)
difference(SE) p <0.001 p <0.001
Bolus insulin dose -1.09 -3.93
(IU/day) change from (0.465) (0.470)
Baseline: LSM difference p = 0.020 p <0.001
(SE)
Effect estimate per Comparison group Sotagliflozin 400 mg
comparison versus placebo
Net benefit (%) (primary Difference 13.4
endpoint)
95% CI 8.97, 17.81

Assessment report
EMA/178275/2019 Page 77/125
Title: A phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study to evaluate the net clinical benefit
of sotagliflozin as adjunct to insulin therapy in type 1 diabetes
Study identifier LX4211.312
P-value <0.001
A1C (%) LSM difference (SE) -0.46 (0.042)
95% CI -0.54, -0.38
P-value <0.001
Body weight (kg) LSM difference (SE) -2.98 (0.166)
95% CI -3.31, -2.66
P-value <0.001
SBP (mm Hg) LSM difference (SE) -3.5 (1.11)
95% CI -5.7, -1.3
P-value 0.002
Bolus insulin dose LSM difference (SE) -2.84 (0.614)
(IU/day)
95% CI -4.05, -1.64
P-value <0.001
Notes The N for individual endpoints and time points may be less than the number randomized to the group.
A1C = glycosylated hemoglobin (HbA1c); CI = confidence interval; IU = international unit; LSM = least squares
mean; mITT = modified intent-to-treat; SBP = systolic blood pressure; SE = standard error

Clinical studies in special populations

No specific clinical studies have been conducted in special populations.

Age 65-74 Age 75-84 Age 85+

(Older subjects (Older subjects (Older subjects
number /total number /total number /total
number) number) number)

Controlled Trials 247 / 3574 18 / 3574 0 / 3574

Non Controlled trials N/A N/A N/A

Analysis performed across trials (pooled analyses)

Subgroup analyses
In order to increase the precision of subgroup analyses on a larger dataset, a pooled analysis (ES1 and
ES2) of A1C was planned.

A1C change by demographics and Baseline characteristics

Overall, treatment effect was statistically significant in favour of sotagliflozin compared to placebo for
each subgroup category in both the ES1 and ES2 pools. Figure 24 and Figure 25 summarize the analysis
of the primary endpoint by demographic and Baseline characteristic subgroups for the ES1 pool (mITT
population) for sotagliflozin 200 mg and 400 mg, respectively.

Assessment report
EMA/178275/2019 Page 78/125
Figure 24 - Forest plot of difference in the LS mean change from Baseline to Week 24 in A1C
(%) of sotagliflozin 200 mg compared to placebo by subgroups for ES1 pool (mITT
population)

Note: The difference between the mean change from Baseline in A1C (%) and 95% CL estimated from MMRM models
are plotted.

Figure 25 - Forest plot of difference in the LS mean change from Baseline to Week 24 in A1C
(%) of sotagliflozin 400 mg compared to placebo by subgroups for ES1 pool (mITT
population)

The effect of sotagliflozin on HbA1c was analysed by BMI <27 kg/m² and ≥27 kg/m² in the pooled
309/310 study dataset, a similar pattern was observed as when analysed by BMI <25 kg/m² and ≥25
kg/m².

Assessment report
EMA/178275/2019 Page 79/125
Supportive study

Study 204

This was a Phase 2, multicenter, randomized double-blind, placebo-controlled, parallel-group study to

evaluate the safety and efficacy of sotagliflozin 400 mg once daily before the first meal of the day in
young-adult patients with T1DM who had poor glycaemic control on their current insulin regimen. Patients
were randomly assigned (1:1 ratio) to receive sotagliflozin 400 mg (2 x 200-mg tablets) or placebo orally
once daily for 12 weeks. Study drug was given in addition to the patient’s insulin regimen. The
randomization was stratified by insulin delivery method (CSII or MDI) and Week -4 A1C (≤10.0%,
>10.0%). The primary efficacy endpoint was the change from Baseline to Week 12 in A1C for sotagliflozin
400 mg versus placebo. Eligible patients were young adults 18 to 30 years of age, with a confirmed
diagnosis of T1DM at least 1 year prior to informed consent, and A1C ≥9.0%.

Patient disposition and demographics

A total of 87 patients were randomized to the study at 14 sites in the US: 44 patients to the placebo group
and 43 patients to the sotagliflozin 400 mg group. Overall, 75 patients (35 on placebo and 40 on
sotagliflozin) completed the study. Of the 85 patients in the mITT population, 45 (52.9%) were female
and 40 (47.1%) male. The mean (SD) age at study entry was 22.3 (3.81) years. Of all patients in the
mITT population, 46 (54.1%) received insulin via CSII and 39 (45.9%) received insulin via MDI; 37
(43.5%) had Week -4 A1C ≤10% and 48 (56.5%) had Week -4 A1C >10%. The mean (SD) age at T1DM
diagnosis was 10.4 (5.84) years. Approximately 90% of all patients were diagnosed with T1DM before 18
years of age. Mean Baseline A1C values were 9.73% for the placebo group and 9.93% for the sotagliflozin
group.

Efficacy results

For the primary endpoint, change from Baseline in A1C at Week 12, both the placebo group and the
sotagliflozin 400 mg group showed a statistically significant LS mean reduction in A1C of ­0.99% and
­1.33%, respectively (both p <0.001). While the placebo-subtracted A1C reduction in the sotagliflozin
400 mg group reached statistical significance at Weeks 3, 6, and 9, the difference at Week 12 (-0.35%)
did not achieve statistical significance. In the 2 subgroups of patients using CSII at Screening and patients
with Screening (Week ­4) A1C ≤10%, sotagliflozin 400 mg demonstrated larger, placebo-subtracted
mean reductions in A1C of ­0.60% and ­0.75%, respectively.

Sotagliflozin demonstrated a change from Baseline at Week 12 in 2-hour PPG following a standardized
Mixed Meal. The LS mean difference in 2-hour PPG between the placebo and sotagliflozin groups was
­56.6 mg/dL (p = 0.001).

Patients in the sotagliflozin group showed numerically greater mean reductions in bolus, basal, and total
insulin doses at Week 12 compared with placebo.

The change from Baseline at Week 12 for various CGM measures of glycaemic measures including percent
time spent in the target blood glucose range was assessed. Patients treated with sotagliflozin achieved an
increase from 33.1% at Baseline to 43.6% at Week 12 in the mean percent time spent in the target blood
glucose range of 70-180 mg/dL as compared with a slight reduction from 33.5% to 32.9% for patients
treated with placebo.

A difference was seen between sotagliflozin and placebo for the change from Baseline to Week 12 in body
weight. The LS mean change from Baseline in body weight to Week 12 was 1.75 kg in the placebo group
(p <0.001) and ­0.62 kg in the sotagliflozin group (p = 0.12). The LS mean difference between the
placebo and sotagliflozin groups in change from Baseline to Week 12 in absolute body weight was ­2.37
kg (p <0.001).

Assessment report
EMA/178275/2019 Page 80/125
2.5.3. Discussion on clinical efficacy

Design and conduct of clinical studies

Three phase 3 studies (studies 309, 310 and 312) provide the main support for the application. In
addition, two dose response studies (studies 202 and 206) were submitted as well as a small supportive
study (204).

Scientific Advice regarding the clinical development in T1DM was provided by the CHMP at several
occasions. The advice given has in all essential aspects been followed except for the recommendation to
specifically study patients with impaired renal function.

Most recommendations received were incorporated into the sotagliflozin Phase 3 T1DM studies:

1. A dose-ranging study (Study 206) was conducted to assess the safety and efficacy of qd dosing of 75
mg, 200 mg, or 400 mg versus placebo in T1DM.

2. The 309 and 310 studies included a 400 mg dose and the lower dose of 200 mg to include more than
1 dose in the Phase 3 clinical studies.

3. An insulin optimization period beginning 6-weeks prior to randomization in the 309 and 310 Phase 3
studies, which continued through the Core treatment period (Week 24) was included. Appropriateness of
titration was assessed by an independent Insulin Dose Monitoring Committee (IDMC) in a blinded fashion.
Insulin adjustment algorithms for pump and Multiple Daily Injections (MDI) (basal-bolus) regimen were
provided to the Investigators as part of the site manual. These could be adapted per Investigator
discretion.

4. Studies 309 and 310 were designed to evaluate the maintenance of efficacy over 12 months.

5. A plan to assess CV safety in the program was developed.

6. Events of Special Interest (EOSI) were included in the final protocols (adjudication of drug-induced
liver injury [DILI], CV events, SH, and DKA [adjudication of DKA was expanded to include adjudication of
metabolic acidosis]).

7. Assessment of effects on markers of bone metabolism and bone density were included.

8. Stratification according to baseline HbA1c was performed.

9. The exclusion criteria for DKA and SH were revised to be less restrictive.

10. A third Phase 3 study (Study 312) was conducted to increase safety exposure to sotagliflozin and
meet ICH E1 guidelines requirements and to study the drug in a more real world setting.

A Follow-up CHMP Scientific Advice (EMEA/H/SA/2417/2/FU/1/2015) was requested in August 2015 to

obtain agreement on the acceptability of the revised T1DM clinical development plan for sotagliflozin. The
revised program was considered generally adequate by the CHMP and Phase 3 study designs and proposal
for DKA monitoring were generally endorsed.

Overall, the recommendations from these meetings have been implemented in the program or were
appropriately discussed in the dossier by the applicant. The CHMP Guideline (CPMP/EXP/1080/00 rev.1)
“Guideline on clinical investigation of medicinal products in the treatment or prevention of diabetes
mellitus” was also taken into account.

Assessment report
EMA/178275/2019 Page 81/125
Study 202 was conducted in patients with T2DM on concomitant metformin treatment. Four different
doses/dosing schedules were evaluated (75 mg, 200 mg, 200 mg BID and 400 mg). There was a clear
dose response effect with regards to reduction in HbA1c and notably there was a numerically larger
treatment effect with the 400 mg dose compared to the 200 mg BID dosing. The dose-response effect was
not as prominent for other endpoints. While this study was conducted in patients with T2DM, the data
formed the basis for the selection of the 200 mg qd and 400 mg qd doses to be further evaluated in the
Phase 3 T1DM studies.

Study 206 investigated the effect of 3 different doses (75 mg, 200 mg and 400 mg) of sotagliflozin in
patients with T1DM. This study was conducted in parallel with the phase 3 studies, in accordance with
advice given, and thus did not inform on the doses to be selected in the phase 3 program. Although a
dose-response relationship was observed for most of the selected endpoints, e.g. UGE, PPG and body
weight, it is noted that the dose-response with regards to HbA1c was less clear than in T2DM patients. The
data from study 206 however support the choice of doses since the 75 mg dose only provided borderline
effects on HbA1c and body weight. Thus the doses selected appear adequate.

The phase 3 studies in the sotagliflozin T1DM program included Studies 309, 310, and 312. Studies 309
and 310 were conducted in North America and in Europe and Israel, respectively. The third phase 3 study,
Study 312, was conducted globally.

Studies 309 and 310 were of similar design and were of adequate size and duration (24+28 weeks) with
the endpoints assessed at week 24. The studies compared two doses of sotagliflozin (200 mg and 400
mg) with placebo on a background medication of optimised insulin treatment. Since no other treatment
than insulin currently is available in the EU, the choice of placebo as comparator is adequate. The insulin
treatment was optimised for six weeks before randomisation. Optimised insulin treatment was to be
continued throughout the study duration. From the start of the insulin optimising period and throughout
the course of the study, the patients SMBG was closely monitored and insulin doses adjusted. An IDMC
reviewed the SMBG and insulin dosing data on an ongoing basis up to week 24 of the study. The measures
taken to ensure optimised insulin treatment appear adequate.

Study 312 was a study of 24 weeks duration and did not include insulin optimisation. In this study, only
the higher sotagliflozin dose, 400 mg, was compared to placebo. Insulin adjustment algorithms were
provided, but there was no IDMC to assess the insulin optimisation. The aim was to provide data from a
setting more close to a “real world” setting as opposed to studies 309/310.

Glycaemic goals and insulin adjustment algorithms were in place for all three studies. For the first meal on
Day 1, patients were instructed to decrease the calculated (or usual) mealtime carbohydrate bolus dose
by 30%. This recommendation was based on data from the phase 2 studies. After the first meal taken
after study drug had been administered on Day 1, insulin doses in all study groups were adapted based on
individual patients' needs and the initial reduction in bolus insulin dose was not expected to destabilise
patients in the insulin only (i.e. placebo) arm.

Studies 309 and 310 also included two substudies, CGM and DEXA, for which it was prespecified that the
analyses should be made on pooled data due to the small number of subject included.

All studies included TIDM patients >18 years of age and with inadequate metabolic control (HbA1c 7.0%
to 11.0 at screening). There was no upper limit for age or BMI. Patients with eGFR>45 could be included.
Patients with a recent history of severe hypoglycaemia or DKA during the last month before screening
were not eligible. The inclusion and exclusion criteria are considered adequate. Inclusion and exclusion
criteria were essentially similar in study 312 compared to studies 309 and 310, with the exception that no
restrictions were made with regards to the type of insulin regimen.

Double-blinded treatment with placebo, sotagliflozin 200 mg (studies 309 and 310) and 400 mg was
provided throughout the entire study duration. HbA1c and FPG data (and PPG data in the CGM study) was

Assessment report
EMA/178275/2019 Page 82/125
masked to the sites, although at week 16 the HbA1c results were unmasked if they exceeded 11%.
Adequate measure were taken to maintain the blind during the studies although it is questioned whether
masking of FPG values would be efficient considering that patients and sites had access to SMBG values.

The primary objective of studies 309 and 310 was to show that sotagliflozin 200 mg and/or 400 mg was
superior to placebo in reducing HbA1c whereas the primary objective of study 312 was to show that
sotagliflozin 400 mg was superior to placebo in the proportion of patients with A1C <7.0% at Week 24 and
no episode of SH and no episode of DKA after Randomization (net benefit). Although the same endpoints
were included in both studies 309/310 and similarly in study 312, the primary endpoint and the
hierarchical order of the secondary endpoints differed. The CGM substudy compared the effect on percent
time spent inside and outside the target glucose ranges and the DEXA substudy compared the effect on
total fat mass. The objectives and endpoints were adequate.

The applicant pooled data from studies 309 and 310, the ES1 pool. Considering the similarities between
studies 309 and 310, this pool is considered adequate.

A second pool (ES2) was made by pooling data for all placebo treated patients and all patients treated
with sotagliflozin 400 mg. This pool is of less interest considering that optimisation of the insulin
treatment differed between studies 309/310 and study 312, which may influence the outcome and make
interpretation difficult.

In summary, the sample size, treatment allocation procedure and statistical methodology for the main
studies are generally considered acceptable.

The primary endpoint in studies 309 and 310 is change of HbA1c after 24 weeks analysed using a mixed
model for repeated measures analysis (MMRM) and targeting a treatment effect as if patients adhered to
treatment until week 24.

The primary efficacy analysis is based on a missing at random (MAR) assumption and it is not possible to
know if this assumption is applicable. Other methods, such as placebo based imputation, may be
considered giving more adequate estimates of efficacy. Analyses using placebo-based imputation were
performed as a sensitivity analyses and results show very similar point estimates to the primary efficacy
analyses.

For the net benefit endpoint in study 312 an approach imputing “failure” for data missing at week 24 was
used. This is a sufficiently conservative method for handling missing data.

Control of the type I error is reasonable using an alpha-split (between active treatment groups) and
sequential testing (of endpoints) in studies 309 and 310, and only sequential testing (of endpoints) in
study 312.

Missing data was a minor problem for the studies 309 and 310 with less than 10% missing values in the
primary endpoint, while missing values in study 312 were interpreted as non-response (ie non-missing).
The primary efficacy analyses of studies 309 and 310 were based on the MAR assumption and it is not
possible to know if this assumption is applicable. However, under the MNAR assumption and placebo
based imputation analysis in studies 309, 310 the point estimates show similar results to the primary
efficacy analysis, which can be regarded as a confirmation of the acceptability of the MAR assumption. In
study 312, the primary analysis was performed in observed cases under an MNAR assumption imputing
missing values as non-responders. Therefore, the primary analysis uses assumptions for missing values
which are acceptable or confirmed by analyses with acceptable assumptions.

While the primary efficacy target of estimation, or estimand, and type of analysis has been common
standard in studies for antidiabetic drugs in recent years, current discussions could suggest that other
targets of estimation could be preferable. The actual adherence to treatment should be reflected in the
target of estimation. Specifically, since patients are not expected to benefit once treatment is

Assessment report
EMA/178275/2019 Page 83/125
discontinued (e.g. due to adverse events) the treatment effect should be estimated based on observed or
modelled data reflecting adherence to treatment as observed in the study. Other important intercurrent
events to consider are the changes to, or introduction of, other medication that will influence HbA1c
values, including use of protocol-defined rescue medication. In T1DM studies the insulin treatment is
adjusted to provide adequate treatment instead of other rescue treatments, and this is not considered an
issue in these studies. Modelling based on data obtained in the placebo group is considered an acceptable
approach to reflect discontinuation from treatment.

The sensitivity analysis provided used observed data as far as available and an imputation method for
missing data imputing data on based observations in the placebo treatment group. Following the ICH E9
(R1, Step 2b) Addendum (EMA/CHMP/ICH/436221/2017), the primary estimand was based on a modified
ITT approach for effect estimation, which can be regarded as a treatment policy estimand. However, as
stated in the addendum, this estimand requires that data is collected after intercurrent events, ideally for
all subjects and visits. Missing data after intercurrent events may compromise the interpretation of the
results as treatment policy estimands. Therefore, the sponsor was asked to provide information 1)
whether the degree of missing data before and after intercurrent events was similar, and 2) to what
degree missing data after intercurrent events may compromise the interpretation of the results. The
sponsor addressed these questions especially in relation to the applied missing-at-random assumption of
the primary analysis model using a MMRM.

The studies were generally well conducted. The amendments made to the study protocols are not
considered to affect the outcome or interpretation of the data in any of the studies. In general, protocol
deviations were few and balanced between groups in all three studies. There was one incident where
ethical conduct of study 310 was compromised. The incident did not impact the safety and well-being of
the subjects, nor the integrity or conduct of the trial. This issue was adequately handled.

Further to this, it was discovered when analysing PK samples in study 310 that in 42 of the 782
randomised patients the plasma PK samples suggested that they had received incorrect study drug. The
errors were distributed over 22 sites in 13 out of 17 countries participating in the trial. The issue was
adequately investigated and handled by the applicant. Sensitivity analyses were performed, excluding all
affected sites. The applicant concludes that the results of these analyses showed no impact on the
primary and secondary efficacy or safety endpoints. This conclusion is endorsed.

Efficacy data and additional analyses

The three phase 3 studies included in total 2 980 patients, 1 231 on placebo, 263 on sotagliflozin 200 mg
and 962 on sotagliflozin 400 mg. In all three studies, completion rate was high (about 90%) and
drop-outs were balanced between groups. In study 312, more patients dropped out due to adverse events
in the sotagliflozin treated group (45) compared to the placebo treated group (16). In studies 309/310, no
patients were excluded from the mITT population and three patients were excluded from the mITT
population in study 312.

Baseline demographic characteristics were balanced within each study. In study 309, which took place in
the US, patients were somewhat older than in studies 310 and 312. A large proportion of subjects (>70%)
included in the studies were overweight. Overweight was more common in study 309 conducted in the
US, than in study 310 conducted in the EU/Israel. Although less common in the European T1DM
population, overweight still is a problem associated with intensive insulin treatment also in this population
and the population included in the study is therefore considered relevant. A higher proportion of patients
in study 309 were using CSII (60%) as their insulin delivery method compared to study 310 (25%) and
study 312 (40%). Baseline characteristics were balanced between treatment groups and comparable
between studies with the exception of HbA1c which was somewhat higher in study 312 (about 8.2%)

Assessment report
EMA/178275/2019 Page 84/125
compared to studies 309/310 (7.6% and 7.7% respectively). Both in study 309 and 310, all treatment
groups showed a similar decrease in HbA1c of about 0.65% after the initial insulin optimising period.

The primary objective for studies 309 and 310 was met as the change from baseline in HbA1c was
statistically significantly larger for both the sotagliflozin 200 mg and 400 mg dose. The treatment
difference was comparable for both dose levels and of moderate magnitude (about -0.4%, p<0.001). As
expected, pooling of data from these studies did not change the outcome. In both studies, there was
virtually no change from baseline HbA1c in the placebo treated groups at week 24. The sensitivity
analysis of the change from Baseline in A1C (%) at Week 24 using the PMM method with control (placebo)
based imputation confirmed the results of the primary analysis. In study 312, change in HbA1c from
baseline was a secondary endpoint. In this study, where insulin treatment was not optimised before the
start of study treatment, the change in HbA1c was larger in both the sotagliflozin and the placebo groups
compared to studies 309/310. The treatment difference was however comparable (-0.46%).

The proportion of patients achieving HbA1c <7.0% with no SH or DKA (net benefit) was the primary
endpoint in study 312. The study met its primary endpoint as the net benefit difference vs placebo was
13.4% which was statistically significant. The main contributing factor was the change in HbA1c. In
studies 309 and 310 a similar pattern was observed, although more patients had HbA1c <7.0% at
baseline than in study 312. A dose-response was observed in study 309 (12% and 22% for the 200 mg
and 400 mg dose respectively), whereas no apparent difference was observed between doses in study
310 (16% and 17% for the 200 mg and 400 mg dose respectively). The treatment differences were
statistically significant for all comparisons. In the ES1 pool, the corresponding outcome was 14.0% for
sotagliflozin 200 mg and 19.5% for sotagliflozin 400 mg. When the net benefit was analysed by its
components (HbA1c < 7.0% without SH or DKA), the findings were consistently in favour of sotagliflozin
treatment. Generally higher proportions were observed with the higher dose. All outcomes were
statistically significant.

PPG was only evaluated in the CGM substudy and pooling of data for this analysis was predefined due to
the low number of subjects participating. The data show that both doses of sotagliflozin resulted in a
statistically significant decrease in PPG, the treatment difference being somewhat larger for the higher
dose (-35 mg/dL and -41 mg/dL for the 200 mg and 400 mg dose respectively). The findings were
consistent with the data from the phase 2 program.

Across the studies, a statistically significantly larger reduction in FPG was observed with sotagliflozin
treatment compared to placebo although the decrease was modest ranging from 9.8 to 25.7 mg/dL.

As already stated, a large proportion of subjects (>70%) included in the studies were overweight. In all
studies, the initial insulin optimising/run-in period was associated with a weight increase in all treatment
groups. Across the three studies, sotagliflozin treatment resulted in consistent weight decrease of -1.7 to
-1.9 kg with the 200 mg dose and -2.3 to -2.7 kg with the 400 mg dose. In the placebo treated groups a
weight gain of +0.2 to +0.8 kg was observed. The treatment difference versus placebo was statistically
significant for all comparisons. The DEXA substudy showed that the decrease in body weight was largely
due to a loss of fat. This is in line with findings for other SGLT2-inhibitors.

Change in bolus insulin dose was a secondary endpoint in all three studies. In all studies the bolus insulin
dose increased in the placebo treated groups (+3.9 to +6.6%) whereas a decrease was observed in the
sotagliflozin treated groups which was larger in the higher dose groups. The change from baseline in the
sotagliflozin treated groups ranged from -2.6 to -10.4% with a LS mean difference from placebo ranging
from -5.7 to -16.4%. The change in basal insulin dose and total insulin dose was also assessed in all three
studies. The pattern for the basal insulin dose was similar to that observed for the bolus doses but the
percent changes were smaller. In the placebo groups, a LS mean increase in the total insulin dose at week
24 ranging from +0.6 to +2.9% was observed whereas the LS mean decrease in total insulin dose
observed in the sotagliflozin treated groups ranged from – 4.3 to -8.9%.

Assessment report
EMA/178275/2019 Page 85/125
The insulin dose adjustments were closely monitored by the IDMC in studies 309/310. For the vast
majority of visits reviewed, the management of dose adjustments were considered to meet standards of
care. This is supported by the data on HbA1c and insulin doses which show that insulin doses slowly
increased in the placebo treated groups for the total study duration and that HbA1c remained stable in
these groups.

Sotagliflozin treatment resulted in a modest placebo-adjusted decrease in SBP. There was no apparent
difference in the treatment effect in the overall population compared to patients with SBP ≥130 mmHg
whereas a slightly larger reduction was observed in patients with SBP ≥140 mmHg.

The aim of the CGM substudy was to investigate the effect of sotagliflozin on blood glucose variability. The
data showed a statistically significant increase in the time spent in the desired blood glucose range
(70-180 mg/dL). This was explained by a decrease in the time spent in the hyperglycaemic blood glucose
range; whereas the time spent in the hypoglycaemic blood glucose range remained unchanged. CSII has
been shown to reduce blood glucose variability per se. Of the patients included, 49.2% were using CSII as
their insulin delivery method. Data from a subgroup analysis by administration method (CSII or MDI)
however showed comparable results for both groups, thus the outcome does not appear to have been
driven by the method of insulin administration.

A graph showing the 24-hour pattern of CGM measurements for placebo, sotagliflozin 200 mg and 400 mg
respectively, was provided. Surprisingly, the most profound glucose-lowering effect of sotagliflozin in
combination with insulin compared to insulin alone was found to occur during the early morning hours. It
can only be speculated that sotagliflozin counteracts what is called the dawn phenomenon by increasing
urinary glucose excretion in response to rising blood glucose levels. Another possible explanation could be
that renal glucose excretion continues during the night. As no food intake compensates for this glucose
loss, blood glucose steadily decreases. The applicant states that the decline in blood glucose values from
12 AM to 6 AM was not associated with a significant increase in hypoglycaemia.

During the rest of the day, blood glucose levels were generally lower with sotagliflozin 400 mg compared
to placebo, whereas the difference between sotagliflozin 200 mg and placebo was less pronounced.

Patient related outcomes were assessed using the DTSQs and DDS2 scores. The DTSQs scores (range 0
to 36) indicated a higher satisfaction with sotagliflozin compared to placebo as reflected by a statistically
significant increase in the score (+1.7 and +2.5 for sotagliflozin 200 mg and 400 mg respectively) as well
as a higher proportion of patients reporting a change in score of 3 or more (54% and 52 % for sotagliflozin
vs 28% for placebo). The DDS2 scores (range 0 to 6) indicated less distress with sotagliflozin treatment
(-0.3 to -0.8, difference from placebo). These data may reflect the benefits of less variable blood glucose
and weight loss observed with sotagliflozin treatment.

No direct comparison between the doses investigated in studies 309 and 310 (200 mg and 400 mg) was
made; instead both doses were compared to placebo. Numerically there was no consistent treatment
difference between the two doses with regards to HbA1c, neither in the overall population nor in the
subgroups analysed. The only exception was patients with BMI >30 kg/m2, where the 400 mg dose
resulted in a larger HbA1c reduction. The higher dose consistently showed a higher risk of DKA except in
patients with BMI >30 kg/m2. Based on these data the use of the higher dose is questioned. However, as
shown by the applicant in the responses to the LoOIs, there was a wide variability in the response with
regards to HbA1c. The difference in treatment effect with regards to reduction in body weight was about
1 kg and the CIs do not overlap. There is a trend towards a better effect on time in range and SBP with the
400 mg dose and the reduction of SH was more pronounced with the 400 mg dose. Taking the individual
response into account, a dose increase could be beneficial in some patients, especially in overweight
patients in need of improved glycaemic control, thus the 400 mg dose is considered justified.

Assessment report
EMA/178275/2019 Page 86/125
One-year exploratory efficacy data was provided from studies 309 and 310. The data show that HbA1c
increased somewhat after week 24, the time point at which the IDMC did no longer review the insulin dose
adjustments. This possibly illustrates the difficulties in maintaining a very tight metabolic control in T1DM.
There was still a modest but statistically significant treatment difference between placebo and
sotagliflozin of about -0.25% for the 200 mg dose and -0.3% for the 400 mg dose. Net benefit had
decreased somewhat at week 52, now ranging from 7.2 to 13.4% as compared to 12 to 22% at week 24.
This was mainly due to a lower proportion of patients achieving HbA1c <7.0%. The body weight slightly
increased up to week 52 in the placebo treated groups, whereas body weight remained stable and
reduced in all sotagliflozin treated groups after week 24 and up to week 52. No apparent change in the
effect on total fat mass was observed. The insulin doses remained essentially unchanged after week 24
with the exception of study 309, where an increase in bolus dose was observed with the 200 mg dose.
Patient reported outcomes measuring diabetes distress remained in favour of sotagliflozin.

No specific clinical studies have been conducted in special populations. The number of older subjects
included in the studies was low, with only 18 subjects aged 75 to 85 years included. Additional analyses
provided with the responses to the LoQ, gave no indication of a different effect due to age per se.
Furthermore, the clinical data in patients with impaired renal function is scarce. It should be noted that
the CHMP recommended that the applicant should consider investigating this population in a dedicated
study. Subgroup analyses by renal function showed no attenuation of the effect in subjects with eGFR ≥60
to <90 mL/min/1.73 m2 compared to patients with eGFR >90 mL/min/1.73 m2. Subjects with eGFR <60
mL/min/1.73 m2 were few, resulting in very wide CIs although the point estimates remained in favour of
sotagliflozin treatment.

Pooled subgroup analyses from studies 309/310 were provided for the two doses, 200 mg and 400 mg
(ES1 pool) as well as for the ES2 pool (all patients treated with sotagliflozin 400 mg). The analyses
showed consistent findings for all demographic and baseline characteristics tested with regards to the
change from baseline in HbA1c, except for patients with eGFR <60 where the 95%CI included zero in all
three analyses. The number of patients were few, thus the 95%CI were wide. Notably, no patients aged
75 years or above were included in these analyses.

A small supportive study (study 204) was also submitted which included 85 young adults (18-30 years
old) with T1DM and a high mean HbA1c of almost 10%, treated with either placebo or sotagliflozin 400
mg. The findings were in line with the outcome of the phase 3 studies.

Additional expert consultations

CHMP requested an ad hoc expert meeting to obtain the opinion of experts in the field of diabetes, as well
as from patient representatives, on the benefits of the use of sotagliflozin in T1DM and the risk of diabetic
ketoacidosis (see section 2.6.1. Discussion on clinical safety).

2.5.4. Conclusions on clinical efficacy

Three well conducted clinical studies of adequate design have been submitted to support the use of
sotagliflozin, a dual SGLT2 and SGLT1 inhibitor in the treatment of T1DM. Treatment with sotagliflozin
when added to optimised insulin treatment resulted in a modest decrease of HbA1c of about 0.4% after
24 weeks. This effect was attenuated after 52 weeks of treatment. A comparable effect was observed
when sotagliflozin was added to a less strictly controlled insulin treatment. There was no apparent
difference between doses with regards to reduction in HbA1c. Beneficial effects were also observed on
PPG, body weight and blood glucose variability. In addition, a higher treatment satisfaction was observed
with sotagliflozin treatment. The data in the elderly (>75 years of age) and in patients with moderate
renal impairment (eGFR <60 mL/min/1.73 m2) is very limited and does not allow any firm conclusions.

Assessment report
EMA/178275/2019 Page 87/125
2.6. Clinical safety

The safety of sotagliflozin was evaluated in a total of 30 clinical studies, including 22 phase 1, 5 phase 2
and 3 phase 3 studies. The phase 2 and 3 studies included subjects with T1DM (3 phase 2 and 3 phase 3
studies) and T2DM (2 phase 2 studies).

Overall, during phase 2 and 3 studies of sotagliflozin in subjects with T1DM, 1,915 subjects were treated
with sotagliflozin of which 35 with sotagliflozin 75 mg, 559 with sotagliflozin 200 mg and 1,321 with
sotagliflozin 400 mg.

For the integrated summary of safety, data was pooled (Table 23):

The SAF-1 pool includes pooled safety data from 2 placebo-controlled phase 3 studies (309 and 310) in
subjects with T1DM with similar study design (insulin optimization and committee oversight of insulin
dosing) for 52 weeks of treatment.

The SAF-3 pool includes pooled safety data from 6 placebo-controlled studies in subjects with T1DM;
3 phase 2 studies (203, 204 and 206) and 3 phase 3 studies (309, 310 and 312). Study 312 was
conducted in subjects with any insulin regimen, less frequent visits and no committee oversight of insulin
dosing. Moreover, study 312 did not include sotagliflozin at the 200 mg dose and had only 24 weeks of
treatment. However, SAF-3 represents the largest pool of data from T1DM subjects.

The SAF-4 pool includes pooled safety data from T1DM and T2DM subjects from 8 placebo-controlled
studies; the 6 studies in T1DM subjects included in SAF-3 (203, 204, 206, 309, 310 and 312) and 2 phase
2 studies in T2DM subjects (201 and 202).
Table 23 Safety pools for clinical study data
Study Pool Description of the studies Number of patients
randomized to treatment

SAF-1 T1DM Sotagliflozin; n=1,049

Phase 3 studies (200 mg; n=524,
(309, 310)
Placebo-controlled, randomized, double-blind 400 mg; n= 525)
52 weeks (24 w+ 28 w) Placebo; n=526

SAF-3 T1DM Sotagliflozin; n=1,915

Phase 2 and 3 studies (75 mg; n=35
(203, 204, 206, 309, 310,
Placebo-controlled, randomized, double-blind 200 mg; n=559
312)
400 mg; n=1,321)
Placebo; n=1,324

SAF-4 T1DM + T2DM Sotagliflozin; n=2,175

Phase 2 and 3 studies (75 mg; n=92
(201, 202, 203, 204, 206,
Placebo-controlled, randomized, double-blind 200 mg; n=631
309, 310, 312)
400 mg; n=1,452)
Placebo; n=1,396

Patient exposure

In total, 1,915 T1DM subjects in phase 2 and 3 studies were treated with sotagliflozin with 1,444 patients
(75%) treated for at least 24 weeks and 668 patients (64%) treated for at least 52 weeks. A total of 260

Assessment report
EMA/178275/2019 Page 88/125
patients with T2DM were exposed to sotagliflozin in the phase 2 and 3 clinical trials. An overview of
exposure in the safety data pools SAF-1, SAF-3 and SAF-4 in Table 24.
Table 24 Overview of treatment exposure for SAF-1, SAF-3 and SAF-4
Safety Pool sotagliflozin Sotagliflozin All sotagliflozin Placebo
200 mg 400 mg

Tot Mean Tot Mean Tot Mean Tot Mean

subj exposure subj exposure subj exposure subj exposure
(Median (Median (Median (Median
exposure) exposure) exposure) exposure)

Patient-yea Patient-yea Patient-year Patient-y

rs (PY) rs (PY) s (PY) ears (PY)

SAF-1 524 334.6 days 525 331.9 days 1,049 333.3 days 526 329.3 days
(364.0 days) (364.0 days) (364.0 days) (364 days)
T1DM
Phase 3
studies 480 PY 477 PY 957 PY 474 PY

SAF-3 *) 559 319.0 days 1,321 219.7 days 1,915 246.1 days 1,324 219.7 days
(364.0 days) (170.0 days) (190.0 days) (169.0
T1DM
days)
Phase 2 and 3
studies 488 PY 795 PY 1,290 PY 796 PY

SAF-4 **) 631 290.7.0 days 1,452 206.6 days 2,175 225.6 days 1,396 211.9 days
(364.0 days) (169.0 days) (172.0 days) (169.0
T1DM +
days)
T2DM
Phase 2 and 3 502 PY 821 PY 1,343 PY 810 PY
studies

*) SAF-3; Included also exposure to 75 mg sotagliflozin (n=35, mean exposure=76.8 days (median expos.=85.0 days);7.36 PY

**) SAF-4; Included also exposure to 75 mg sotagliflozin (n=92, mean exposure=77.5 days (median expos.=84.0 days);19.53 PY

Adverse events

An overall summary of adverse events for SAF-1 through 52 weeks of treatment is presented in
Table 25.

In SAF-1, about 75% of the subjects reported AEs. The frequency of SAEs was about 10% for sotagliflozin
and 7% for placebo. The discontinuation rates due to adverse events were slightly higher for sotagliflozin
(4-7%) compared to placebo (4%).

Assessment report
EMA/178275/2019 Page 89/125
Table 25 Overall summary of TEAEs (treatment-emergent adverse events) through 52 weeks
of treatment in SAF-1
Placebo Sotagliflozin Sotagliflozin All
200 mg 400 mg sotagliflozin
N =526 N = 524 N =525 N = 1,049
Patients with TEAEs n (%) n (%) n (%) n (%)
Patients with any TEAE 374 (71.1) 393 (75.0) 390 (74.3) 783 (74.6)
Patients with treatment-related TEAEs 106 (20.2) 167 (31.9) 193 (36.8) 360 (34.3)
Patients with severe TEAEs 37 (7.0) 50 (9.5) 48 (9.1) 98 (9.3)
Patients with severe treatment-related TEAEs 11 (2.1) 19 (3.6) 22 (4.2) 41 (3.9)
Patients with treatment-emergent SAEs 37 (7.0) 53 (10.1) 50 (9.5) 103 (9.8)
Patients with treatment-emergent 10 (1.9) 18 (3.4) 23 (4.4) 41 (3.9)
treatment-related SAEs
Patients with TEAEs leading to study drug 20 (3.8) 23 (4.4) 35 (6.7) 58 (5.5)
discontinuation
Patients with treatment-related TEAEs leading to 12 (2.3) 19 (3.6) 31 (5.9) 50 (4.8)
study drug discontinuation
Patients with any TEAEs leading to death 3 (0.6) 0 0 0

Most frequently reported adverse events

In SAF-1, the most frequently reported adverse events for sotagliflozin 200 mg and 400 mg versus
placebo were viral upper respiratory tract infection (15% and 14% vs. 13%), diarrhoea (6.5% and 9.3%
vs. 5.5%), diabetic ketoacidosis (4.0% and 5.7% vs. 1.0%), blood ketone body increased (4.0% and
5.1% vs. 0.6%), urinary tract infections (6.1% and 3.8% vs. 4.9%), genital infection fungal (3.4% and
4.0% vs. 0.2%) and vulvovaginal mycotic infection (2.9% and 5.1% vs. 1.9%), Table 26.
Table 26 Most common TEAEs (reported by ≥2% of patients in either sotagliflozin group and
more frequently than placebo) through 52 weeks of treatment in SAF-1
Placebo Sotagliflozin Sotagliflozin All
200 mg 400 mg sotagliflozin
System organ class (N = 526) (N = 524) (N = 525) (N = 1,049)
Preferred term n (%) n (%) n (%) n (%)
Patients with any common TEAE 374 (71.1) 393 (75.0) 390 (74.3) 783 (74.6)
Ear and labyrinth disorders 3 (0.6) 6 (1.1) 11 (2.1) 17 (1.6)

Eye disorders 19 (3.6) 22 (4.2) 20 (3.8) 42 (4.0)

Gastrointestinal Disorders 97 (18.4) 106 (20.2) 122 (23.2) 228 (21.7)

Diarrhoea 29 (5.5) 34 (6.5) 49 (9.3) 83 (7.9)
Vomiting 15 (2.9) 16 (3.1) 11 (2.1) 27 (2.6)

Constipation 9 (1.7) 17 (3.2) 6 (1.1) 23 (2.2)

Flatulence 0 5 (1.0) 14 (2.7) 19 (1.8)

Abdominal pain 3 (0.6) 3 (0.6) 11 (2.1) 14 (1.3)

General Disorders and Administration 40 (7.6) 47 (9.0) 48 (9.1) 95 (9.1)

Site Conditions
Pyrexia 10 (1.9) 10 (1.9) 13 (2.5) 23 (2.2)
Fatigue 10 (1.9) 8 (1.5) 11 (2.1) 19 (1.8)
Infections and Infestations 233 (44.3) 241 (46.0) 243 (46.3) 484 (46.1)
Viral upper respiratory tract infection 69 (13.1) 78 (14.9) 74 (14.1) 152 (14.5)
Urinary tract infection 26 (4.9) 32 (6.1) 20 (3.8) 52 (5.0)
Vulvovaginal mycotic infection 10 (1.9) 15 (2.9) 27 (5.1) 42 (4.0)
Genital infection fungal 1 (0.2) 18 (3.4) 21 (4.0) 39 (3.7)
Gastroenteritis 12 (2.3) 15 (2.9) 17 (3.2) 32 (3.1)
Sinusitis 16 (3.0) 15 (2.9) 16 (3.0) 31 (3.0)
Gastroenteritis viral 7 (1.3) 16 (3.1) 8 (1.5) 24 (2.3)
Investigations 47 (8.9) 64 (12.2) 66 (12.6) 130 (12.4)
Blood ketone body increased 3 (0.6) 21 (4.0) 27 (5.1) 48 (4.6)

Assessment report
EMA/178275/2019 Page 90/125
 Placebo Sotagliflozin Sotagliflozin All
200 mg 400 mg sotagliflozin
System organ class (N = 526) (N = 524) (N = 525) (N = 1,049)
Preferred term n (%) n (%) n (%) n (%)
Metabolism and Nutrition Disorders 41 (7.8) 64 (12.2) 82 (15.6) 146 (13.9)
Diabetic ketoacidosis 5 (1.0) 21 (4.0) 30 (5.7) 51 (4.9)
Hypoglycaemia 12 (2.3) 12 (2.3) 9 (1.7) 21 (2.0)
Ketosis 3 (0.6) 4 (0.8) 17 (3.2) 21 (2.0)
Vitamin D deficiency 5 (1.0) 10 (1.9) 11 (2.1) 21 (2.0)
Musculoskeletal and Connective Tissue 74 (14.1) 75 (14.3) 82 (15.6) 157 (15.0)
Disorders
Back pain 11 (2.1) 14 (2.7) 20 (3.8) 34 (3.2)
Arthralgia 9 (1.7) 16 (3.1) 11 (2.1) 27 (2.6)
Neoplasms benign, malignant and 8 (1.5) 10 (1.9) 13 (2.5) 23 (2.2)
unspecified

Nervous system disorders 66 (12.5) 58 (11.1) 56 (10.7) 114 (10.9)

Headache 19 (3.6) 15 (2.9) 20 (3.8) 35 (3.3)
Dizziness 9 (1.7) 13 (2.5) 9 (1.7) 22 (2.1)
Renal and Urinary Disorders 22 (4.2) 37 (7.1) 35 (6.7) 72 (6.9)
Pollakiuria 7 (1.3) 16 (3.1) 10 (1.9) 26 (2.5)
Reproductive system and breast 13 (2.5) 19 (3.6) 25 (4.8) 44 (4.2)
disorders

Respiratory, Thoracic, and Mediastinal 59 (11.2) 35 (6.7) 56 (10.7) 91 (8.7)

Disorders
Cough 12 (2.3) 12 (2.3) 25 (4.8) 37 (3.5)
Oropharyngeal pain 14 (2.7) 4 (0.8) 17 (3.2) 21 (2.0)

Adverse events of special interest

Diabetic ketoacidosis

During every visit, Investigators were trained to review patient reported signs or symptoms that might be
suspicious for diabetic ketoacidosis (DKA). In addition, patients were instructed by the Investigator to
measure blood or urine ketones or blood BHB level. If the urine ketones were positive or blood BHB level
was >0.6 mmol/L, the patient was asked to contact the investigative site immediately. Under these
circumstances, the patient was asked to increase hydration and administer additional rapid acting insulin
with oral carbohydrate frequently, as often as every 2 hours until normalization of urine ketones or blood
BHB level.

In SAF-1, there was a dose-dependent increase in the incidence of positively adjudicated DKA for
sotagliflozin (2.9% and 3.8% for sotagliflozin 200 mg and 400 mg, respectively) compared to placebo
(0.2%). The exposure-adjusted incidence rate (EAIR) was 3.12, 4.19 and 0.21 subjects per 100 PY for
sotagliflozin 200 mg, sotagliflozin 400 mg and placebo, respectively. In total, 37 events of DKA in 36
subjects were received (of which 26 events of DKA for sotagliflozin). One subject, in the sotagliflozin
group 200 mg, experienced more than one event of DKA. All events of DKA were serious. The overall
incidence of events of DKA leading to study discontinuation was 33%[27% (4/15) for sotagliflozin 200
mg, 40% for sotagliflozin 400 mg (8/20) and 0% (0/1) for placebo] and the overall incidence resulting in
study interruption was 56% [67% (10/15) for sotagliflozin 200 mg, 45% for sotagliflozin 400 mg (9/20)
and 100% for placebo (1/1)] (Table 27). In SAF-1, 43% (15/35) of the cases experienced DKA with
glucose values in the euglycaemic range (<14 mmol/L).

Assessment report
EMA/178275/2019 Page 91/125
Table 27 Summary of treatment-emergent Investigator-reported and positively adjudicated
acidosis-related events in SAF-1
Placebo Sotagliflozin Sotagliflozin All sotagliflozin
(N = 526) 200 mg 400 mg (N = 1,049)
n (%) m (N = 524) (N = 525) n (%) m
n (%) m n (%) m
Total subject-years
474.20 PY 480.02 PY 477.13 PY 957.15 PY

Investigator-reported DKA/metabolic acidosis events

All events 7 ( 1.3) 7 30 ( 5.7) 32 39 ( 7.4) 39 69 ( 6.6) 71

Positively adjudicated metabolic acidosis events

All events 3 ( 0.6) 3 18 ( 3.4) 19 22 ( 4.2) 22 40 ( 3.8) 41

Positively adjudicated DKA events

All events 1 ( 0.2) 1 15 ( 2.9) 16 20 ( 3.8) 20 35 ( 3.3) 36

Events per patient per year 0.002 0.033 0.042 0.038

EAIR per 1000 patient years 2.11 31.25 41.92 36.57

(95% CL) ( 0.00,6.24) ( 15.43,47.06) ( 23.55,60.29) ( 24.45,48.68)

Severe events 0 11 ( 2.1) 12 15 ( 2.9) 15 26 ( 2.5) 27

Serious events 1 ( 0.2) 1 15 ( 2.9) 16 20 ( 3.8) 20 35 ( 3.3) 36

Events leading to study drug 1 ( 0.2) 1 10 ( 1.9) 10 9 ( 1.7) 9 19 ( 1.8) 19

interruption
Events leading to study drug 0 4 ( 0.8) 4 8 ( 1.5) 8 12 ( 1.1) 12
discontinuation
m = number of events, EAIR=Exposure-adjusted incidence rate Source: modified from 5.3.5.3 ISS Part 2 Table 1.10.3.1

In SAF-3 (excluding study 203), the exposure-adjusted incidence rate was higher in the sotagliflozin
groups (3.07 and 5.29 subjects per 100 PY for sotagliflozin 200 mg and 400 mg, respectively) compared
to placebo (0.76 subject per 100 PY), also in a dose-dependent manner. In total, 64 events of DKA in 63
subjects were received (of which 58 events of DKA for sotagliflozin). One subject, in the sotagliflozin
group 200 mg, experienced more than one event of DKA. All but 3 positively adjudicated DKA were
serious and all but 5 DKA led to hospitalization. Overall, the incidence of events of DKA leading to study
discontinuation and study interruption was 38% and 46%, respectively.

The incidence of positively adjudicated DKA increased over time for sotagliflozin as compared to placebo
in a clearly dose-dependent manner. In SAF-1, the cumulative incidence of DKA at week 52 was
approximately 2.8% and 4.1% for sotagliflozin 200 mg and 400 mg sotagliflozin compared to 0.2% for
placebo (Figure 26). The Kaplan Meier plot of the cumulative incidence of adjudicated DKA in SAF-3
showed a more pronounced differentiation between the 200 mg and 400 mg dose of sotagliflozin, with
earlier TTO with 400 mg compared to the SAF-1 plot. This may reflect the data from study 312 where no
insulin optimization occurred prior to randomization (Figure 27).

In SAF-1, mean and median time to onset of DKA was shorter in sotagliflozin 400 mg (159 and 134 days)
than in sotagliflozin 200 mg (208 and 214 days); although SD standard deviation was broad the
differences suggests a shorter TTO with the sotagliflozin 400 mg compared to sotagliflozin 200 mg.

Assessment report
EMA/178275/2019 Page 92/125
Figure 26 Kaplan-Meier plot of the cumulative incidence of positively adjudicated diabetic
ketoacidosis over time by treatment group SAF-1

Figure 27 Kaplan-Meier plot of the cumulative incidence of positively adjudicated diabetic

ketoacidosis over time by treatment group SAF-3

Possible risk factors/contributing factors associated with an event of a positively adjudicated DKA were
analysed. However, the analyses of potential risk factors of DKA were associated with uncertainties due to
low number of events. However, the subgroup analyses indicated that subjects with previous DKA,
subjects with high BHB levels at baseline or greater increase in BHB levels, subjects who had their insulin
dose decreased with more than 20% and subjects who were CSII-users, were at higher risk to develop
DKA. This has been reflected in the proposed product information.

The applicant has proposed risk minimisation measures in order to mitigate the risk of DKA which includes
education of health care professionals and patients about self-monitoring of ketones, situations at-risk to
allow for early diagnosis of ketonaemia or ketonuria and how to manage potential ketosis and
maintenance of optimal insulin therapy. Detailed recommendations are provided in the SmPC and in the
proposed educational materials. These recommendations are considered adequate in order to mitigate
the risk, based on the knowledge on DKA in general and on the experience from the clinical trials.

Assessment report
EMA/178275/2019 Page 93/125
Hypoglycaemia

In SAF-1, the incidence of documented hypoglycaemia was high across all treatment groups (>98%). The
event rate per patient-year was similar in the sotagliflozin groups (81 and 84 events per PY for
sotagliflozin 200 mg and 400 mg, respectively) and slightly lower compared to placebo (96 events per
PY), Table 28. The incidence of recurrent documented hypoglycaemia events was similar across
treatment groups in subjects having 1-5, 6-9 or >10 events.

The incidence of events of positively adjudicated severe hypoglycaemia was slightly higher for
sotagliflozin 200 mg (5.7%) compared to sotagliflozin 400 mg (4.4%) but was even higher for placebo
(7.4%). The event rate was higher for sotagliflozin 200 mg (142 events per 1000 PY) compared to
sotagliflozin 400 mg (69 events per 1000 PY) and placebo (105 events per 1000 PY). The majority of
patients had one or two events of severe hypoglycaemia; however, subjects having ≥3 events of SH
occurred in 30% (10/30) of the subjects in the sotagliflozin 200 mg group. In the Kaplan-Meier plot of the
cumulative incidence of severe hypoglycaemia, the sotagliflozin curves became less steep over time and
provided a separation from placebo after about 24 weeks of treatment (Figure 28).

Overall, 7 subjects discontinued due to events of severe hypoglycaemia in the phase 2 and phase 3 pooled
studies; 1 subject in the sotagliflozin 75 mg group, 1 subject in the sotagliflozin 200 mg, 2 subjects in the
sotagliflozin 400 mg and 3 subjects in the placebo group.
Table 28 Summary of events of documented hypoglycaemia and events of positively
adjudicated severe hypoglycaemia through 52 weeks of treatment in SAF-1
Placebo Sotagliflozin Sotagliflozin All
200 mg 400 mg sotagliflozin
(N = 526) (N = 524) (N = 525) (N = 1,049)

Total patient-year 474.20 480.02 477.13 957.15

Incidence of documented hypoglycaemia

45,327 39,015 39,937 78,952

Number of events of documented hypoglycaemia

Patients with at least 1 event, n (%) 518 ( 98.5) 515 ( 98.3) 518 ( 98.7) 1,033 ( 98.5)

EAIR per 1,000 patient-years 1092.37 1072.87 1085.65 1079.24

Events per patient per year 95.587 81.278 83.702 82.486

Number of nocturnal events 5,771 5,286 5,313 10,599

Number of diurnal events 39,766 33,889 34,796 68,685

Incidence of positively adjudicated severe hypoglycaemia

Number of events of positively adjudicated events 50 68 33 101

of severe hypoglycaemia

Patients with at least 1 event, n (%) 39 (7.4) 30 (5.7) 23 (4.4) 53 (5.1)

EAIR per 1000 patient-years 82.24 62.50 48.20 55.37

Events per patient per year 0.11 0.14 0.07 0.11

Number of events nocturnal by time of day 13 20 6 26

Number of events nocturnal by sleep status 17 24 15 39

Number of events diurnal by time of day 36 48 29 77

Assessment report
EMA/178275/2019 Page 94/125
Figure 28 Kaplan-Meier plot of the cumulative incidence of positively adjudicated severe
hypoglycaemia over time by treatment group SAF-1 (52-week Phase 3 Studies) - Safety
population

Volume depletion

Due to the mechanism of action, sotagliflozin induces an osmotic diuresis, which may lead to adverse
events related to volume depletion. In the SAF-3 Pool, the incidence rate of events of volume depletion
was higher in sotagliflozin-treated subjects (29 and 26 subjects per 1000 PY for sotagliflozin 200 mg and
400 mg, respectively) compared to placebo-treated subjects (10 subjects per 1000 PY).

In SAF-1, the incidence of osmotic diuresis-related adverse events (‘urine output increased’, ‘polydipsia’,
‘micturition urgency’, ‘nocturia’, ‘pollakiuria’, ‘polyuria’) was increased for sotagliflozin relative to placebo.
The incidence of volume depletion events (‘dehydration’, ‘hypovolaemia’, ‘postural dizziness’, ‘orthostatic
hypotension’, ‘hypotension’, ‘syncope’, ‘presyncope’) was 2.7%, 1.1% and 1.0% for sotagliflozin 200 mg,
sotagliflozin 400 mg and placebo. Increased urination and volume depletion have been reflected in the
proposed SmPC.

In the subgroup analyses in the SAF-3 Pool, subjects ≥65 years of age (n=209) had a higher incidence of
events of volume depletion for sotagliflozin 200 mg and 400 mg relative to placebo. Subjects with eGFR
<60 mL/min/1.73 m2 (n=146) had a higher incidence of events of volume depletion for sotagliflozin 400
mg versus placebo. However, the total number of subjects in each analysis was small.

Genital mycotic infections

In SAF-1, the incidence of genital infections was higher in patients treated with sotagliflozin compared to
placebo patients, and the increase was dose-related in both male and female patients.

In SAF-1, the incidence of genital infections in female subjects was 15%, 17% and 4.7% in sotagliflozin
200 mg, sotagliflozin 400 mg group and placebo. Most of the events were mild or moderate and no serious
case was reported. Discontinuation due to genital mycotic infections occurred in 1.2%, 1.1% and 0.8% of
subjects treated with sotagliflozin 200 mg, sotagliflozin 400 mg and placebo, respectively. Eighty-seven
(87) female subjects experienced 130 events of genital infections. Recurrent events of genital infections
were reported in 27% (35/130) of the female subjects in the sotagliflozin group in SAF-3.

Assessment report
EMA/178275/2019 Page 95/125
In SAF-1, the incidence of male genital infections was 3.0%, 6.3% and 1.1% in sotagliflozin 200 mg,
sotagliflozin 400 mg group and placebo. All events were mild or moderate in intensity and no serious
cases. Discontinuation due to genital mycotic infections occurred in 0%, 0.4% and 0.4% of subjects
treated with sotagliflozin 200 mg, sotagliflozin 400 mg and placebo, respectively. Twenty-four (24) male
subjects experienced 29 events of genital infections. Recurrent events of genital infections were reported
in 13% (4/31) of the male subjects in the sotagliflozin group in SAF-3.

Urinary tract infections

In SAF-1, the overall frequency of urinary tract infections reported were slightly higher for sotagliflozin
200 mg (7.1%) compared to sotagliflozin 400 mg (5.5%) and placebo (6.1%). The incidence of UTI in
female subjects was highly increased and similar in subjects treated with sotagliflozin 200 mg (12%) and
placebo (11%) compared to sotagliflozin 400 mg (7.0%); however, the incidence of UTI in male subjects
was dose-dependently increased in the sotagliflozin 200 mg and 400 mg group (2.3% and 4.0%,
respectively) as compared to placebo (1.8%). In SAF-3, all UTI events were mild or moderate in intensity
except for one severe case (male subject in the sotagliflozin 400 mg group). Two cases
(2 cases of cystitis) were serious; both occurred in male subjects in the sotagliflozin 400 mg group.
Information regarding urinary tract infections has been reflected in the proposed SmPC.

Diarrhoea

In SAF-1, the incidence of diarrhoea was increased in the sotagliflozin 200 mg and 400 mg group (6.5%
and 8.8%, respectively) as compared to placebo (5.1%). In SAF-3, all events were mild or moderate in
intensity except for one severe case in the sotagliflozin 200 mg group. No serious cases and few subjects
(<1%) in all groups) had diarrhoea events that lead do discontinuation.

Renal events

In SAF-1, treatment with sotagliflozin was associated with a small decrease in eGFR; -3.6 and -3.8
mL/min/1.73 m2 (-4.0% and -4.3%) in sotagliflozin 200 mg and sotagliflozin 400 mg vs. -1.15
mL/min/1.73 m2 (-1.3%) in the placebo group that that did return towards baseline at week 52 and week
53 (1 week after discontinuation of study drug), Figure 29. Also in subjects with moderate and mild renal
impairment, the eGFR returned towards baseline at week 52. However, subjects with eGFR <60
mL/min/1.73 m2 made up for a small number in SAF-1 (n=71), wherefore difficult to draw any firm
conclusions in this subgroup. Mean increases in serum creatinine from baseline to week 4 was 3.0, 3.2
and 1.0 umol/L (4.0%, 4.3% and 1.3%) for sotagliflozin 200 mg, sotagliflozin 400 mg and placebo,
respectively. The changes in eGFR and serum creatinine have been reflected in the proposed SmPC.

The incidence of renal-related events was low and similar across the sotagliflozin groups and placebo in
both SAF-1 (1.5%, 1.3% and 1.5%) and SAF-3 (16, 16 and 15 subjects per 1000 PY). One patient each
in the placebo group, sotagliflozin 200 mg group, and sotagliflozin 400 mg group experienced a serious
renal event.

In the subgroup analysis in the SAF-3 Pool, subjects with eGFR ≥90 mL/min/1.73 m2 had a higher
incidence of renal events for sotagliflozin (9.3 and 16 subjects per 1000 PY for sotagliflozin 200 mg and
400 mg, respectively) relative to placebo (7.5 subjects per 1000 PY). In subjects with eGFR 60 to <90
mL/min/1.73 m2, the incidence of subjects with renal events was higher for sotagliflozin 200 mg
(20 subjects per 1000 PY) compared to placebo (14 subjects per 1000 PY) and sotagliflozin 400 mg (11
subjects per 1000 PY). Subjects with eGFR <60 mL/min/1.73 m2 (n=146) had a higher incidence of renal
events in in the placebo group (103 subjects per 1000 PY) compared to sotagliflozin 200 mg and 400 mg
(45 and 84 subjects per 1000 PY); however the total number of subjects was small.

Assessment report
EMA/178275/2019 Page 96/125
Figure 29 Mean Change (+/-SE) in Estimated Glomerular Filtration Rate (eGFR) mL/min/1.73
m2) From Baseline Over Time, 52 Weeks of Treatment SAF-1

Hepatic events

In the phase 2 and 3 studies in T1DM and T2DM subjects (SAF-4), the incidence of positively adjudicated
drug-induced liver damage (DILI) was low; 1 event in the placebo group and 2 events in the sotagliflozin
400 mg group. No case met the definition for Hy’s law case.

Pancreatitis

One case of pancreatitis was identified in a patient randomized to sotagliflozin 400 mg. The event of
pancreatitis resolved one week after onset. The pancreatitis was during a case of DKA.

Venous thrombotic events

One patient randomized to sotagliflozin 400 mg, with a history of thrombophlebitis, was reported to have
experienced 2 events of VTE (pulmonary embolism and deep vein thrombosis).

Hypersensitivity reactions

In the phase 2 and 3 studies in T1DM and T2DM subjects (SAF-4), sotagliflozin treatment did not result in
a higher incidence of potential severe hypersensitivity reactions relative to placebo. The incidence of
potential severe events of hypersensitivity reactions was low and similar in the sotagliflozin 200 mg and
400 mg groups (0.5% respectively) and the placebo group (0.7%).

Bone safety

In sub-studies of studies 309 and 310, DEXA assessments of bone density were performed at baseline
and week 52. A total of 243 patients were randomized in the pooled DEXA sub-study, and 215 patients
completed the DEXA sub-study; 75 patients in the placebo group, 84 patients in the sotagliflozin group
200 mg and 84 subjects in the sotagliflozin 400 mg.

The incidence of fractures through 52 weeks of treatment was low and similar across the groups: 2.9%,
1.9% and 3.4% for sotagliflozin 200 mg, sotagliflozin 400 mg and placebo. Additional studies to
characterize the risk of bone fractures for sotagliflozin are ongoing in T2DM subjects; data should be
provided when available.

Assessment report
EMA/178275/2019 Page 97/125
The change from baseline to week 52 in BMD was nominally statistically significant in the ‘lumbar spine’
for sotagliflozin 200 mg [LS Mean difference from placebo -1.27% (-2.51%, -0.03%), p=0.044] and
sotagliflozin 400 mg [LS Mean difference from placebo -2.53% (-3.79%, -1.28%), p<0.001] and in the
‘total hip’ for sotagliflozin 400 mg [LS Mean difference from placebo -1.18% (-2.07%, -1.28%),
p=0.009].

At week 52, a decrease in 1.25-dihydroxyvitamin D (-16% and -19% vs. -3.0%) and an increase in
25-hydroxyvitamin D (40% and 41% vs. 34%) was observed for sotagliflozin 200 mg and 400 mg vs.
placebo. A slight increase in phosphate for sotagliflozin 200 mg (8.7%) and sotagliflozin 400 mg (1.5%)
vs. placebo (-2.5%) and a slight and similar increase in PTH for sotagliflozin 400 mg (4.2%) and placebo
(3.3%) vs. sotagliflozin 200 mg (-0.83%) and no meaningful change in calcium
(0.81% and -0.51% vs. -0.28%) was noted for sotagliflozin 200 mg and 400 mg vs. placebo. However, all
changes in mean concentrations were within the reference range at baseline through week 52.

There was a dose-dependent increase from baseline to week 52 in the bone resorption marker CTX for
sotagliflozin 200 mg and 400 mg (18% and 32%) relative to placebo (1.2%) and a non-dose-dependent
increase in P1NP (8.7% and 3.4% vs. -4.9% for sotagliflozin 200 mg and 400 mg vs. placebo). There are
plausible theories which indicate that treatment with SGLT2-inhibitors could induce decreases in BMD and
increases in bone turnover caused by weight reduction. Long-term clinical relevance of the increase in
CTX with regards to safety is unknown. ‘Bone fractures’ is included as an important potential risk in the
RMP and will be closely monitored in ongoing clinical trials.

Cardiovascular risk

In the phase 2 and 3 studies in T1DM and T2DM subjects (SAF-4), the incidence rate of positively
adjudicated MACE was higher for sotagliflozin 200 mg (9.97 subjects per 1000 PY) than for sotagliflozin
400 mg (4.95 subjects per 1000 PY) and placebo (8.65 subjects per 1000 PY); however, the incidence of
MACE was lower in the ‘all sotagliflozin’ group (6.88 subjects per 1000 PY) compared to placebo. The
estimated hazard ratio for positively adjudicated MACE in the ‘all sotagliflozin’ group relative to placebo
was 0.68 and the upper CI 1.82. There was no notable difference in restricted mean survival time values
between the all sotagliflozin group and placebo. Considering the known mechanism of action and
experience from other products in the class, the provided data is reassuring although no firm conclusions
could be drawn.

Malignancies

The incidence of malignancies ‘of special interest’ was low and similar across the groups (0.1-0.3%).
When using the SMQ Malignant or unspecified tumours, there was no imbalance in the incidence of
malignancies (0.4-0.6%). Further analysis, to identify subjects reporting a malignancy with onset greater
than 6 months after the first dose of study medication, did also show low and similar incidence of
malignancies across the groups (0.1-0.3%). Malignancy reported in more than one subject in the
sotagliflozin groups was breast cancer (1 intraductal proliferative breast lesion and 1 invasive breast
carcinoma). No trend could be observed, although, the risk for developing malignancies cannot be fully
explored from controlled data in the clinical trial program covering rather short observation periods
(mean duration less than a year).

Lower limb amputations

In total, there were two events of toe amputations (of which one transmetatarsal) in the sotagliflozin 200
mg and 400 mg group, respectively. Both cases had medical histories of multiple toe amputation. In
accordance to the finalised CHMP/PRAC Referral procedure, a class warning regarding the increased risk
of limb amputations has been included in the proposed SmPC.

Assessment report
EMA/178275/2019 Page 98/125
Serious adverse events and deaths

Deaths

A total of 4 deaths occurred in the clinical studies, of which 3 in the placebo group and 1 in the sotagliflozin
group (completed suicide).

Non-fatal serious adverse events

Non-fatal SAEs were most frequently reported in the SOC Metabolism and nutrition disorders, of which
DKA was the most commonly reported event (3.6%, 5.0% and 0.6% for sotagliflozin 200 mg, sotagliflozin
400 mg and placebo).

Laboratory findings

Haematology

A slight increase from baseline to week 52 in haematocrit concentration was observed for sotagliflozin
(2%) compared to placebo (0%). The proportion of subjects that met the PCI criteria (haematocrit
>50%) were higher in the sotagliflozin 200 mg and 400 mg groups (6.7% and 8.2%, respectively)
compared to the placebo group (2.7%). The observed increase in haematocrit is considered related to
volume depletion associated with diuretic effect of sotagliflozin, as for other SGLT2 inhibitors. Increased
haematocrit has been reflected in the proposed SmPC.

Serum lipids

Mean percent changes from baseline for sotagliflozin 200 mg and 400 mg versus placebo were: HDL-C
3.3% and 4.2% vs. 0.5%; LDL-C 5.0% and 6.1% vs. 3.3%; triglycerides 5.7% and 5.4% vs. 2.7%.
Serum lipid changes have been reflected in the proposed SmPC.

Beta hydroxybutyrate (BHB)

BHB values that met the PCI criteria of ≥0.6 mmol/L were higher in the sotagliflozin 200 mg and 400 mg
groups (47% and 46%, respectively) compared to the placebo group (20%) and BHB values that met the
PCI criteria of ≥1.0 mmol/L were 20%, 18% and 6.8% for sotagliflozin 200 mg, sotagliflozin 400 mg and
placebo, respectively.

Liver function tests

The percentages of subjects with increases in ALT and AST that met PCI criteria >3x ULN were similar
across the groups or slightly higher in the placebo group; with increases in ALT (0.8%, 1.3% and 1.5%)
and AST (1.3%, 1.3% and 1.5%) for sotagliflozin 200 mg, sotagliflozin 400 mg and placebo, respectively.
The portion of subjects with increases in ALT or AST that met PCI criteria >5x ULN were low (0.4%
increases in ALT across all groups and 0.4-0.6% increases in AST for sotagliflozin and 0% for placebo). No
subjects had liver function test values that met Hy’s Law criteria.

Blood pressure/ pulse rate

Slight changes in in SBP and DBP were observed with sotagliflozin vs. placebo (mean changes of -1.6,
-2.2 and 0.9 mmHg in SBP and -1.4, -1.0 and 0.1 mmHg in DBP for sotagliflozin 200 mg, sotagliflozin 400
mg and placebo). No clinically relevant mean change from baseline in pulse rate was observed.

Assessment report
EMA/178275/2019 Page 99/125
Safety in special populations

Elderly

Overall in the clinical studies, there is rather limited data in subjects ≥65 years (n=265) and very limited
data in subjects ≥75 years (n=18). Initiation of sotagliflozin in subjects above 75 years of age is not
recommended. ‘Use in subjects ≥75 years’ has been included as missing information in the proposed
RMP.

Gender

In SAF-1, adverse events were in general slightly more common in females (75-78%) than in males
(68-73%) across the groups.

Race/Ethnicity

In SAF-1, adverse events were in general more common in non-whites (72-97%) than in whites
(69-75%) across the groups; however, there is rather limited data for non-whites (n=92).

Baseline (week-2) A1C

The overall incidences of AEs were similar or slightly higher across the groups in subjects with baseline
A1C ≤8.5% compared to subjects with baseline A1C >8.5%. The overall incidences of SAEs were higher
across the groups in subjects with baseline A1C ≤8.5% (8-11%) compared to subjects with baseline A1C
>8.5% (3-8%).

Body mass index

In SAF-1, the overall incidences of AEs and SAEs were similar or slightly higher across the groups in
subjects with a BMI <25 kg/m2 compared to subjects with a BMI ≥ 25 kg/m2.

Subjects with renal impairment

In SAF-1, the subgroup of patients with eGFR <60 mL/min/1.73 m2 made up for a small amount of
subjects (n=71) wherefore difficult to draw any conclusions. In subjects with eGFR ≥60 to <90
mL/min/1.73 m2 (n=774), the incidence of AEs and SAEs were similar across the groups; however, the
incidence of volume depletion was slightly higher in the sotagliflozin groups (2.6% and 1.2% in
sotagliflozin 200 mg and 400 mg, respectively) compared to placebo (0.8%). In subjects with eGFR ≥90
mL/min/1.73 m2 (n=730), the overall incidence of AEs and SAEs were higher in sotagliflozin 200 mg
(72% and 9.9%) and sotagliflozin 400 mg (75% and 8.3%) compared to placebo (68.9% and 4.7%).

In subgroup analyses in the SAF-3 Pool, subjects with eGFR <60 mL/min/1.73 m2 (n=146) had a higher
incidence of events of volume depletion for sotagliflozin 400 mg versus placebo; however the total
number of subjects was small.

Subjects with hepatic impairment

Safety has not been established in subjects with moderate or severe hepatic impairment.

Immunological events

N/A

Assessment report
EMA/178275/2019 Page 100/125
Safety related to drug-drug interactions and other interactions

Several phase 1 clinical pharmacology studies were conducted in healthy subjects to evaluate potential
interaction between sotagliflozin and drugs commonly used in the T1DM population. No safety signal was
identified in these studies.

Sotagliflozin is intended to be prescribed with insulin, and its impact on insulin dosing has been well
characterized in phase 2 and 3 studies.

Discontinuation due to AEs

In SAF-1, the discontinuation rates due to AEs were higher in the sotagliflozin groups (4.4% and 6.7% for
sotagliflozin 200 mg and 400 mg, respectively) than in the placebo group (3.8%). The most frequently
reported AEs leading to study drug discontinuation was in the SOC Metabolism and nutrition disorders, of
which DKA was the most commonly reported event (0.8%, 1.9% and 0% for sotagliflozin 200 mg,
sotagliflozin 400 mg and placebo).

2.6.1. Discussion on clinical safety

The database is in general considered sufficient since a substantial number of subjects have been
included. In total, 1,915 T1DM subjects were treated with sotagliflozin for a total exposure of 1,290 PY
with 1,444 subjects (75%) treated for at least 24 weeks and 668 subjects (64%) treated for at least 52
weeks.

The approach for integrating and presenting safety data is acceptable. The primary safety pool (SAF-1)
included two 52 weeks phase 3 studies in T1DM subjects. SAF-1 allows for comparison between placebo-
and sotagliflozin-treated groups (200 mg and 400 mg). The two studies were conducted in subjects
treated in the same way, i.e. with insulin optimization and committee oversight of insulin dosing. The
broad safety pool (SAF-3) included T1DM subjects from both phase 2 and 3 studies (the two 52 weeks
studies in SAF-1 and a world-wide 24 weeks study of sotagliflozin 400 mg vs placebo in T1DM subjects
with insulin therapy that was not optimized). SAF-3 represents the largest pool of data from T1DM
subjects.

Overall, the safety database is considered to be representative of a broad general population of adults
with T1DM and includes a sufficient number of subjects with long duration of disease. However, there is
rather limited data in subjects ≥65 years (n=265) and very limited data in subjects ≥75 years (n=18).

Discontinuation rates were similar across the treatment groups in SAF-1 (about 15%). The most common
reason for discontinuation from study drug was withdrawal by subject and discontinuation due to adverse
events.

In SAF-1, about 75% of the subjects reported AEs. The most frequently reported AEs for sotagliflozin 200
mg and sotagliflozin 400 mg versus placebo were viral upper respiratory tract infection (15% and 14% vs.
13%), diarrhoea (6.5% and 9.3% vs. 5.5%), diabetic ketoacidosis (4.0% and 5.7% vs. 1.0%), blood
ketone body increased (4.0% and 5.1% vs. 0.6%), urinary tract infections (6.1% and 3.8% vs. 4.9%),
genital infection fungal (3.4% and 4.0% vs. 0.2%) and vulvovaginal mycotic infection (2.9% and 5.1%
vs. 1.9%).

Diabetic ketoacidosis

In SAF-1, there was a dose-dependent increase in the incidence of positively adjudicated DKA for
sotagliflozin (2.9% and 3.8% for sotagliflozin 200 mg and 400 mg, respectively) compared to placebo
(0.2%). The exposure-adjusted incidence rate was 3.12, 4.19 and 0.21 subjects per 100 PY for
sotagliflozin 200 mg, sotagliflozin 400 mg and placebo, respectively. In total, 37 events of DKA in 36

Assessment report
EMA/178275/2019 Page 101/125
subjects were received (of which 26 events of DKA for sotagliflozin). One subject (in the sotagliflozin
group) experienced more than one event of DKA. All events of DKA were serious. The overall incidence of
events of DKA leading to study discontinuation and study interruption was 33% and 56%, respectively.

In SAF-3 (excluding study 203), the exposure-adjusted incidence rate for positively adjudicated DKA was
higher in the sotagliflozin groups (3.07 and 5.29 per 100 PY for sotagliflozin 200 mg and 400 mg,
respectively) compared to placebo (0.76 per 100 PY), also in a dose-dependent manner. In total, 64
events of DKA in 63 subjects were received (of which 58 events of DKA for sotagliflozin).

The incidence of positively adjudicated DKA increased over time for sotagliflozin as compared to placebo
in a clearly dose-dependent manner. In SAF-1, the cumulative incidence of DKA at week 52 was
approximately 2.8% and 4.1% for sotagliflozin 200 mg and 400 mg sotagliflozin compared to 0.2% for
placebo. The Kaplan Meier plot of the cumulative incidence of adjudicated DKA in SAF-3 showed a more
pronounced differentiation between the 200 mg and 400 mg dose of sotagliflozin, with earlier TTO with
400 mg compared to the SAF-1 plot. This may reflect the data from study 312 where no insulin
optimization occurred prior to randomization.

Mean time to onset of DKA was shorter in sotagliflozin 400 mg (159 days) than in sotagliflozin 200 mg
(208 days); although SD standard deviation was broad the difference suggests a shorter TTO with the
sotagliflozin 400 mg compared to sotagliflozin 200 mg.

Possible risk factors/contributing factors associated with an event of a positively adjudicated DKA was
analysed. However, the analyses of potential risk factors of DKA were associated with uncertainties due to
low number of events. However, the subgroup analyses indicated that subjects with previous DKA,
subjects with high BHB levels at baseline or greater increase in BHB levels, subjects who had their insulin
dose decreased with more than 20% and subjects who were CSII-users, were at higher risk to develop
DKA. This has been reflected in the proposed product information. In 19% (7/35) of the cases, no cause
(no contributing factor/ risk factor) to the DKA event could be identified.

The applicant has proposed risk minimisation measures in order to mitigate the risk of DKA which includes
education of health care professionals and patients about self-monitoring of ketones, situations at-risk to
allow for early diagnosis of ketonaemia or ketonuria and how to manage potential ketosis and
maintenance of optimal insulin therapy. Detailed recommendations are provided in the SmPC and in the
proposed educational materials. These recommendations are considered adequate in order to mitigate
the risk, based on the knowledge on DKA in general and on the experience from the clinical trials.

Hypoglycaemia

In SAF-1, the incidence of documented hypoglycaemia was high across all treatment groups (≥98%). and
the event rate was similar in the sotagliflozin groups (81 and 84 events per PY, respectively) and higher
in the placebo group (96 events per PY). The incidence of recurrent documented hypoglycaemia events
was similar across treatment groups in subjects having 1-5, 6-9 or >10 events.

The incidence of events of positively adjudicated severe hypoglycaemia was slightly higher for
sotagliflozin 200 mg (5.7%) compared to sotagliflozin 400 mg (4.4%) but was even higher for placebo
(7.4%). In the Kaplan-Meier plot of the cumulative incidence of severe hypoglycaemia, the sotagliflozin
curves became less steep over time and provided a separation from placebo after about 24 weeks of
treatment.

Volume depletion

In SAF-1, the incidence of volume depletion was 2.7%, 1.1% and 1.0% in sotagliflozin 200 mg,
sotagliflozin 400 mg and placebo.

Assessment report
EMA/178275/2019 Page 102/125
In the subgroup analyses in the SAF-3 Pool, subjects ≥65 years of age (n=209) had a higher incidence of
events of volume depletion for sotagliflozin 200 mg and 400 mg relative to placebo. Subjects with eGFR
<60 mL/min/1.73 m2 (n=146) had a higher incidence of events of volume depletion for sotagliflozin 400
mg versus placebo. However, the total number of subjects in each analysis was small.

Genital mycotic infections

There was an increased risk of genital infections in sotagliflozin-treated subjects relative to placebo. Both
female (15% and 17% vs 4.7%) and male (3.0% and 6.3% vs 1.1%) genital infections were highly
increased compared to placebo. Most of the events were mild to moderate in intensity.

Urinary tract infections

The overall frequency of urinary tract infections was slightly higher for sotagliflozin 200 mg (7.1%)
compared to sotagliflozin 400 mg (5.5%) and placebo (6.1%). The incidence of UTI in female subjects
was highly increased and similar in subjects treated with sotagliflozin 200 mg (12%) and placebo (11%)
compared to sotagliflozin 400 mg (7.0%); however, the incidence of UTI in male subjects was
dose-dependently increased in the sotagliflozin 200 mg and 400 mg group (2.3% and 4.0%, respectively)
as compared to placebo (1.8%).

Diarrhoea

The incidence of diarrhoea was increased in the sotagliflozin 200 mg and 400 mg group (6.5% and 8.8%,
respectively) as compared to placebo (5.1%).

Renal events

Treatment with sotagliflozin was associated with a small decrease in eGFR in the sotagliflozin 200 mg and
400 mg groups (-4.0% and -4.3%) compared to placebo (-1.3%) that did return towards baseline at
week 52. Also in subjects with moderate and mild renal impairment, the eGFR returned towards baseline
at week 52. However, subjects with eGFR <60 mL/min/1.73 m2 made up for a small number (n=71) in
SAF-1, wherefore difficult to draw any firm conclusions in this subgroup.

The incidence of renal-related events was low and similar across the sotagliflozin groups (1.5% and
1.3%) and placebo (1.5%).

In the subgroup analysis in the SAF-3 Pool, subjects with eGFR 60 to <90 mL/min/1.73 m2 had a higher
incidence of renal events in sotagliflozin 200 mg (20 subjects per 1000 PY) compared to placebo (14
subjects per 1000 PY) and sotagliflozin 400 mg (11 subjects per 1000 PY). Subjects with eGFR <60
mL/min/1.73 m2 (n=146) had a higher incidence of renal events in in the placebo group (103 subjects per
1000 PY) compared to sotagliflozin 200 mg and 400 mg (45 and 84 subjects per 1000 PY); however the
total number of subjects was small to draw any firm conclusions.

Bone safety

The incidence of fractures through 52 weeks of treatment in study 309 and 310 was low and similar across
the groups: 2.9%, 1.9% and 3.4% for sotagliflozin 200 mg, sotagliflozin 400 mg and placebo. Additional
studies to characterize the risk of bone fractures for sotagliflozin are ongoing in T2DM subjects and data
should be provided when available.

The change from baseline to week 52 in BMD was nominally statistically significant in the ‘lumbar spine’
for sotagliflozin 200 mg [LS Mean difference from placebo -1.27 (-2.51, -0.03), p=0.044] and
sotagliflozin 400 mg [LS Mean difference from placebo -2.53 (-3.79, -1.28), p<0.001] and in the ‘total
hip’ for sotagliflozin 400 mg [LS Mean difference from placebo -1.18 (-2.07, -1.28), p=0.009].

An increase in phosphate and 25-hydroxyvitamin D, a decrease in 1.25-dihydroxyvitamin D and no

meaningful change in calcium were observed for sotagliflozin relative placebo. A slight and similar

Assessment report
EMA/178275/2019 Page 103/125
increase in PTH for sotagliflozin 400 mg (4.2%) and placebo (3.3%) vs. sotagliflozin 200 mg (-0.83%)
was noted. However, all changes in mean concentrations were within the reference range at baseline
through week 52.

There was a dose-dependent increase from baseline to week 52 in the bone resorption marker, CTX, for
sotagliflozin 200 mg and 400 mg (18% and 32%) relative to placebo (1.2%) and a non-dose-dependent
increase in the bone formation marker, P1NP (8.7% and 3.4% vs. -4.9% for sotagliflozin 200 mg and 400
mg vs. placebo). Long-term clinical relevance of clinical relevance of the increase in CTX with regards to
safety is unknown. ‘Bone fractures’ is included as an important potential risk in the RMP and will be closely
monitored in ongoing clinical trials.

Cardiovascular risk

The estimated hazard ratio for positively adjudicated MACE in the all sotagliflozin group relative to
placebo was 0.68 and the upper CI 1.82. There was no notable difference in restricted mean survival time
values between the all sotagliflozin group and placebo. Considering the known mechanism of action and
experience from other products in the class, the provided data are reassuring although no firm
conclusions could be drawn.

Malignancies

There was no imbalance in the incidence of malignancies (0.4-0.6%). Further analysis, to identify
subjects reporting a malignancy with onset greater than 6 months after the first dose of study medication,
did also show low and similar incidence of malignancies across the groups (0.1-0.3%). Malignancy
reported in more than one subject in the sotagliflozin groups was breast cancer (1 intraductal proliferative
breast lesion and 1 invasive breast carcinoma). No trend could be observed, although, the risk for
developing malignancies cannot be fully explored from controlled data in the clinical trial program
covering rather short observation periods (mean duration less than a year).

Lower limb amputations

In total, there were two events of toe amputations (of which one transmetatarsal) in the sotagliflozin 200
mg and 400 mg group, respectively. Both cases had medical histories of multiple toe amputation.

Laboratory findings

Slight increases from baseline to week 52 in hematocrit (1.9% and 2.0% vs. 0%) and in serum lipids
(HDL-C: 3.3% and 4.2% versus 0.5%; LDL-C: 5.0% and 6.1% versus 3.3%; triglycerides: 5.7% and
5.4% versus 2.7%) were noted for sotagliflozin 200 mg and 400 mg versus placebo.

Subgroups

In subgroup analyses in the SAF-3 Pool, there was rather limited data for subjects ≥65 years of age and
subjects with eGFR <60 mL/min/1.73 m2 to draw any firm conclusions.

In subjects ≥65 years of age (n=209), there was an increased risk of events related to volume depletion
for sotagliflozin (102 and 38 EAIR per 1000 PY for sotagliflozin 200 mg and 400 mg, respectively)
compared to placebo (zero events).

In subjects with eGFR 60 to <90mL/min/1.73 m2 (n=1,492), the incidence of events of volume depletion
was higher for sotagliflozin (28 and 24 subjects per 1000 PY for sotagliflozin 200 mg and 400 mg)
compared to placebo (8.3 subjects per 1000 PY). In subjects with eGFR <60 mL/min/1.73 m2 (n=146),
the incidence of volume depletion was 111 subjects per 1000 PY for sotagliflozin 400 mg and 0 events for
sotagliflozin 200 mg and placebo, respectively. Sotagliflozin should not be initiated in subjects with
moderate renal impairment (eGFR <60 mL/min).

Assessment report
EMA/178275/2019 Page 104/125
Overall in the clinical studies, there is rather limited data in subjects ≥65 years (n=265) and very limited
data in subjects ≥75 years (n=18). Initiation of sotagliflozin treatment in subjects above 75 years of age
is not recommended. ‘Use in subjects ≥75 years’ has been included as missing information in the
proposed RMP.

Additional expert consultations

CHMP requested an ad hoc expert meeting to obtain the opinion of experts in the field of diabetes, as well
as from patient representatives, on the benefits of the use of sotagliflozin in T1DM and the risk of diabetic
ketoacidosis. Questions were addressed to the ad hoc expert group. The corresponding answers are
presented below:

Question 1

What is the AHEG opinion on the clinical relevance of the treatment effects observed with
sotagliflozin, esp. with regard to reduction of HbA1c, insulin doses, body weight and glucose
variability?

Efficacy outcomes from studies 309, 310 and 312 were presented to the experts. With regard to individual
aspects they had the following view:

The absolute reductions of HbA1c in the treatment groups (by 0.39 to 0.49 % (mean absolute change
from baseline) in studies 309 and 310) were seen as positive and by several experts also as clinical
relevant. Some experts with unrestricted participation agreed with the latter while others considered the
reduction only of borderline clinical relevance.

The experts did not see an obvious benefit in the observed reduction of the average insulin doses (both
basal and post prandial); one expert pointed out that insulin treatment in patients with type 1 diabetes
mellitus (T1DM) is a hormone replacement therapy, therefore, conceptually, a lowering of the insulin dose
may not be a goal at all (at least in non-overweight patients).

The reduction of body weight (by 1.68 to 2.72 kg (mean absolute change from baseline) in studies 309
and 310) was considered of minor benefit, but nevertheless was considered to be beneficial for some
patients, according to some of the experts.

The studies also demonstrated a reduction of the variability of plasma glucose. This was seen by the
experts in general as a relevant goal, in particular as it did not result in an increase of hypoglycaemic
events. The AHEG noted that results from 2 questionnaires (DTSQ and DDS2) did demonstrate patients’
preference for this treatment option and therefore could be supportive of the benefit of reduced glucose
variability seen in the study. The experts found it, however, difficult to judge the clinical relevance of such
patient reported outcomes. A benefit, mentioned by the patients participating in the expert meeting, is
that such an improvement helps to simplify insulin dose calculations. An additional benefit noted was the
lowering of blood pressure, although this was deemed relevant only by some of the experts.

The AHEG had a split view whether the totality of the efficacy outcomes demonstrated in studies would
represent a clinically important benefit overall for patients with T1DM. Also among the experts with
unrestricted participation some considered this to be clinically important while others as being of
borderline clinical relevance.

Considering this to be a potentially lifelong treatment for T1DM, and the lack of data for sotagliflozin in
T2DM (though available for already approved SGLT-2 inhibitors), experts and patients saw a need for the
generation of more data with long-term treatment, including renal outcomes, and safety outcomes in
general.

Question 2

Assessment report
EMA/178275/2019 Page 105/125
The risk of DKA was considerably higher in the sotagliflozin treated groups compared to
placebo despite careful information to patients and monitoring including measurements of
ketones.

a. Please discuss the acceptability of this risk in clinical practice

All experts agreed that DKA represents a substantial and important risk. The experts also noted that the
incidence of DKA in the real world is considerably higher than the incidence seen in the control group of
the studies, presumably due to the highly selected patient population and study sites, as well as measures
implemented during the trial to mitigate this risk. Nevertheless, the experts noted that cases of DKA
where reported with similar frequency at all time points during the studies.

The experts stated that education and awareness of the problem, including the occurrence of DKA with
only slightly elevated plasma glucose levels, is of foremost importance to reduce the risk.

The experts acknowledged the focused approach by the applicant to mitigate the risk, and thought
measures such as 1-2 weeks intensified training prior to start of medication as beneficial.

The AHEG had split views on the acceptability of this risk. A majority of experts considered that the risk
might be manageable in a restricted subpopulation (in particular in patients well trained and
well-educated in treatment of T1DM). Some experts with unrestricted participation agreed with the latter
while others were of the opinion that the expected extent of DKA could constitute an unacceptable risk in
clinical practice.

The risk of DKA with the use of sotagliflozin maybe different in different health care systems as, according
to experts, once available, the product could be expected to be prescribed also by less well-trained
generalists in some member states.

b. Risk minimisation measures such as a guide for health care professionals and
patients, as well as a Patient Alert Card have been proposed in order to mitigate the
risks. What is the AHEG view on the appropriateness and effectiveness of the proposed
measures in clinical practice?

The experts, including the 2 patients, raised some concerns that the proposed monitoring and
precautionary activities may put quite a high burden on health care professionals and responsibility on
patients.

The AHEG noted that the proposed minimisation of the risk of DKA relies heavily on the measurement of
ketones, both before initiating therapy with sotagliflozin and in particular in case of signs or symptoms of
DKA as well as circumstances perceived as posing an increased risk of DKA. It was acknowledged that
ketone measurements by patients, as outlined by the applicant, represent an important contribution to
the safe use of the product. It was pointed out, however, that this was not easily available in some
member states, and even if this was the case, whether this was practical to measure on a frequent base
and would achieve high acceptance by patients was questioned. Also, experience with this seems to
indicate that slight increases for various causes (e.g. diet) may cause frequent follow-ups by health care
professionals. From the patients’ perspective the acceptance of another frequently to be self-measured
laboratory parameter beyond glucose was questioned.

c. Please discuss any additional potential measures that could be introduced to decrease
the risk of DKA.

The AHEG emphasized that general awareness and education of DKA (in particular also euglycaemic
DKA), both with patients and health care professionals, is of high importance to reduce the risk. There

Assessment report
EMA/178275/2019 Page 106/125
was also the unanimous view that to mitigate the risk treatment and prescription of this therapy should be
exclusively by specialists.

The experts had no further proposals for risk mitigation.

Question 3

Please discuss which patients, if any, with T1DM could relevantly benefit from treatment with
sotagliflozin, i.e. what could be a potential target population in clinical practice? Could
restrictions with respect to BMI and insulin requirements be of relevance?

According to the applicant, the population should be selective, e.g. to include only patients who are
compliant, who are willing and able to perform the foreseen ketone testing scheme, who do not have high
alcohol consumption, who do not intend to get pregnant etc. The experts found any benefit in patients
with T1DM could be expected to be most relevant in patients who are overweight, who are well educated
and trained in T1DM, and who have large glucose variability. While it was acknowledged that the risk of
DKA seemingly is higher in patients treated with insulin pump therapy, it was also said that those patients
may particularly benefit as often suffering from high glucose variability in the first place, and in any case
represent an increasing and important proportion of the T1DM population.

The experts were asked to discuss possible parameters indicative of an improved benefit/risk ratio. The
total daily insulin dose as one possible parameter was discussed. While it was acknowledged that patients
with low insulin requirements may be of somewhat increased risk of DKA, the experts were sceptical of
the daily insulin dose as a possible parameter, as under some circumstances even patients with low
insulin requirements may benefit and because insulin requirements may be the consequence of other
underlying circumstances which should be taken into account. The experts thought that a low HbA1c
value should not constitute an exclusion criterion per se, but that highly elevated HbA1c should, also
considering that the latter may reflect a low degree of patient compliance. The experts also thought that
body weight could be a relevant parameter, as the safety profile would improve with a higher BMI. One
expert emphasized the importance of C-peptide levels, as an important marker of residual beta-cell
function relevant for definition of the most appropriate target population to define an indication.

The experts agreed that the general T1DM population is in any case a too broad target population. They
further agreed that proper patient education and compliance is very essential and, if approved for T1DM,
the product should exclusively be prescribed in specialist centres, or at least by specialists, i.e.
diabetologists or endocrinologists.

2.6.2. Conclusions on clinical safety

The safety profile for sotagliflozin is in general expected and most consistent with other SGLT2 inhibitors
regarding the increased risk of volume depletion, genital infections and UTI. Inhibition of SGLT1 by
sotagliflozin delays and reduces glucose absorption in the proximal intestine, which can lead to diarrhoea.

Initiation of sotagliflozin treatment in subjects above 75 years of age is not recommended. Sotagliflozin
should not be initiated in subjects with moderate renal impairment (eGFR <60 mL/min).

The major safety concern of sotagliflozin is the increased and dose-dependent risk of DKA when used in
T1DM patients. The increased risk is of concern considering that DKA is a condition which is potentially
life-threatening that require hospitalisation. The incidence rate of DKA in T1DM patients is estimated to be
about 3% (Hamdy and Khardori 2014) up to 7% (Weinstock et al 2013, Maahs et al 2015).

Analyses of potential risk factors of DKA were associated with uncertainties due to low number of events.
However, the subgroup analyses showed that subjects with previous DKA, subjects with high BHB levels
at baseline or greater increase in BHB levels, subjects who had their insulin dose decreased with more

Assessment report
EMA/178275/2019 Page 107/125
than 20% and subjects who were CSII-users, were at higher risk to develop DKA. This has been reflected
in the proposed product information.

The applicant has proposed risk minimisation measures in order to mitigate the risk of DKA which includes
education of health care professionals and patients about self-monitoring of ketones, situations at-risk to
allow for early diagnosis of ketonaemia or ketonuria and how to manage potential ketosis and
maintenance of optimal insulin therapy. Detailed recommendations are provided in the SmPC and in the
proposed educational materials. These recommendations are considered adequate in order to mitigate
the risk, based on the knowledge on DKA in general and on the experience from the clinical trials.

2.7. Risk Management Plan

Safety concerns

Pharmacovigilance plan

Study Summary of Safety concerns Milestones Due dates

Status objectives addressed

Risk of DKA in To evaluate the DKA Protocol 2 months after

T1DM in patients incidence of DKA submission approval
treated with with sotagliflozin Registration in 2 months after
sotagliflozin as an treated patients as the EU PAS approval
adjunct to insulin compared to non register
versus insulin alone SGLTi treated
patients to Jul-2021;
Planned Dec-2021;
determine Interim report
Category 1 effectiveness of Jul-2022;
additional risk Dec-2022;
minimization Jul-2023;
measures for DKA Dec-2023;
in the post approval Jul-2024 (To be
setting adjusted based on
launch plans)
Final report Dec-2024 (To be
adjusted based on
launch plans)
DKA: Diabetic Ketoacidosis; EU: European Union; PAS: Post-Authorization Study; SGLTi: Sodium–Glucose
Co-transporter inhibitors; T1DM: Type 1 Diabetes Mellitus.

Study Summary of Safety concerns Milestones Due dates

Status objectives addressed

Assessment report
EMA/178275/2019 Page 108/125
 Study Summary of Safety concerns Milestones Due dates
Status objectives addressed

A randomized The safety database of Lower limb CSR Q3 2022

double-blind, this study will provide amputation
placebo-controlle safety information. Pancreatitis
d, parallel-group,
multicenter study Malignancies
to demonstrate Bone fractures
the effects of Long-term
sotagliflozin on cardiovascular
cardiovascular safety
and renal events
Patients ≥75 years
in patients with
type 2 diabetes,
cardiovascular
risk factors and
moderately
impaired renal
function
Ongoing
Category 3

A 26-week The safety database of Bone fractures First Step Q2 2020

randomized, this study will provide Analysis CSR
double-blind, information on bone (6-month
placebo-controlle safety. data)
d, parallel-group, Q4 2021
Second Step
multicenter, Analysis CSR
phase 3 study (24-month
with a 78-week data)
extension period
to evaluate the
efficacy and
safety of
sotagliflozin in
patients 55 years
and older with
T2DM and
inadequate
glycemic control
Ongoing
Category 3

Use of To examine if there is an Malignancies Study 3 months following

Sotagliflozin and association between protocol the drug approval
Risk of sotagliflozin use and the and prior to study
Malignancies in risk of bladder, renal, start
Adult Patients breast, Leydig cell,
First interim 36 months
with Type 1 pancreatic, thyroid, and
report post-launch
Diabetes Mellitus prostate cancers
Second 60 months
Planned
interim post-launch
Category 3 report 84 months
Third interim post-launch
report 108 months
Fourth post-launch
interim 132 months
report post-launch
Final study
report

CSR: Clinical Study Report; DKA: Diabetic Ketoacidosis; EU: European Union; HCP: Healthcare
Professional; PAS: Post-Authorization Study; Q: Quarter; T1DM: Type 1 Diabetes Mellitus; T2DM:
Type 2 Diabetes Mellitus.

Assessment report
EMA/178275/2019 Page 109/125
Risk minimisation measures

Assessment report
EMA/178275/2019 Page 110/125
Conclusion

The CHMP and PRAC considered that the risk management plan version 1.4 is acceptable.

2.8. Pharmacovigilance

Pharmacovigilance system

The CHMP considered that the pharmacovigilance system summary submitted by the applicant fulfils the
requirements of Article 8(3) of Directive 2001/83/EC.

Periodic Safety Update Reports submission requirements

The requirements for submission of periodic safety update reports for this medicinal product are set out
in the Annex II, Section C of the CHMP Opinion. The applicant did not request alignment of the PSUR cycle
with the international birth date (IBD). The new EURD list entry will therefore use the EBD to determine
the forthcoming Data Lock Points.

2.9. New Active Substance

The applicant compared the structure of sotagliflozin with active substances contained in authorised
medicinal products in the European Union and declared that it is not a salt, ester, ether, isomer, mixture
of isomers, complex or derivative of any of them.

The CHMP, based on the available data, considers sotagliflozin to be a new active substance as it is not a
constituent of a medicinal product previously authorised within the European Union.

Assessment report
EMA/178275/2019 Page 111/125
2.10. Product information

2.10.1. User consultation

The results of the user consultation with target patient groups on the package leaflet submitted by the
applicant show that the package leaflet meets the criteria for readability as set out in the Guideline on the
readability of the label and package leaflet of medicinal products for human use.

2.10.2. Quick Response (QR) code

A request to include a QR code in the labelling and package leaflet for the purpose of providing the patient
alert card and educational material (Carer’s guide) has been submitted by the applicant and has been
found acceptable.

2.10.3. Additional monitoring

Pursuant to Article 23(1) of Regulation No (EU) 726/2004, Zynquista (sotagliflozin) is included in the
additional monitoring list as:

• It contains a new active substance which, on 1 January 2011, was not contained in any medicinal
product authorised in the EU.

Therefore the summary of product characteristics and the package leaflet includes a statement that this
medicinal product is subject to additional monitoring and that this will allow quick identification of new
safety information. The statement is preceded by an inverted equilateral black triangle.

3. Benefit-Risk Balance

3.1. Therapeutic Context

3.1.1. Disease or condition

The indication proposed for Zynquista is:

“Zynquista is indicated as an adjunct to insulin therapy to improve glycaemic control in adults with type
1 diabetes mellitus with a Body Mass Index (BMI) ≥ 27 kg/m2, who have failed to achieve adequate
glycaemic control despite optimal insulin therapy.”

Type 1 diabetes mellitus is an autoimmune disease characterized by rapidly progressive pancreatic β-cell
destruction leading to a state of absolute insulin deficiency. The aim of insulin treatment is to substitute
for the lack of endogenous insulin.

3.1.2. Available therapies and unmet medical need

Therapeutic options for type 1 diabetes are limited to intensive insulin or insulin analogue therapy
delivered subcutaneously by MDI or insulin pump. Based largely on the results of the Diabetes Control and
Complications Trial (DCCT), current treatment guidelines recommend that patients with T1DM be treated
to a goal HbA1c of <7.0% with intensive insulin therapy. Despite significant advances in insulin therapies,
delivery methods and management, 70% of T1DM patients do not achieve optimal glycaemic control and
are at increased risk of microvascular complications caused by chronic hyperglycaemia. Furthermore,
some patients with T1DM experience wide fluctuation in their level of glucose control on currently

Assessment report
EMA/178275/2019 Page 112/125
available insulin regimens. Intensive insulin therapy is also associated with an increased risk of
hypoglycaemia, which can be life-threatening and may prevent patients from achieving optimal glycaemic
control. Furthermore, intensive insulin therapy is associated with excessive weight gain and with
peripheral insulin resistance, both being risk factors for hypertension, and CV disease. Approximately
60% of adults living with T1DM are overweight or obese, 60% have dyslipidaemia and 40% have
hypertension, and are therefore at risk of developing macrovascular complications.

During insulin therapy, DKA may occur in situations where insulin administration is interrupted or when
the insulin need is increased, e.g. with severe infections. In the literature, the incidence of DKA in T1DM
patients shows large variations with a range from 0-56 per 1000 person-years. The prevalence of DKA
decreases with increasing age.

For patients with T1DM no other treatment than insulin is authorized in the EU. The fact that most patients
fail to achieve glycaemic targets and are at risk of hypoglycaemia and excessive weight gain on current
type 1 treatments constitute an unmet need that could potentially be addressed by a new treatment used
as an adjunct to insulin.

3.1.3. Main clinical studies

Three Phase 3 studies were included in the sotagliflozin T1DM program (studies 309, 310, and 312).

Studies 309 and 310 had the same design. Both studies were multicenter, randomized, double-blind,
placebo-controlled, parallel-group studies to evaluate the efficacy, safety, and tolerability of sotagliflozin
200 mg and 400 mg versus placebo administered qd as an adjunct therapy in adults with T1DM who had
inadequate glycaemic control with insulin therapy. For each study, the total duration of exposure to study
drug was up to 52 weeks. A key design feature for studies 309 and 310 was insulin optimization prior to
initiation of treatment with sotagliflozin. Study 309 included 793 patients and study 310 included 782
patients.

Study 312 was a Phase 3, multicenter, randomized, double-blind, placebo-controlled, parallel-group

study to evaluate the net benefit (defined as the proportion of patients with A1C <7.0% and no episode
of SH and no episode of DKA from randomization to Week 24) of sotagliflozin 400 mg versus placebo
administered qd as an adjunct to insulin in patients with T1DM. This study did not utilize insulin
optimization, and treatment was not confined to any specific insulin regimen. The total duration of study
participation was up to 32 weeks. The study included 1405 patients.

3.2. Favourable effects

The primary objective for studies 309 and 310 was met as sotagliflozin was shown to be superior to
placebo with regards to the primary endpoint - change in HbA1c from baseline. The treatment difference
for sotagliflozin 200 mg vs placebo was -0.36% and -0.37% (p<0.001) and -0.41% and -0.35%
(p<0.001) for the 400 mg dose in studies 309 and 310 respectively. Pooling of data from these studies did
not change the outcome (-0.41% and -0.43% for 200 mg and 400 mg respectively). In both studies,
there was virtually no change from baseline HbA1c in the placebo treated groups at week 24. In study
312, where change in HbA1c was a secondary endpoint and insulin treatment was not optimised, the
change in HbA1c was larger in both the sotagliflozin and the placebo groups compared to studies
309/310. The treatment difference vs placebo was however comparable (-0.46%; p<0.001).

The proportion of patients achieving HbA1c <7.0% with no SH or DKA (net benefit) was the primary
endpoint in study 312. The study met its primary endpoint as the net benefit difference vs placebo was
13.4% (95% CI 9.12, 17.67) which was statistically significant. The main contributing component was the
proportion of patients achieving HbA1c <7.0% (29.6% vs 15.8% for 400 mg and placebo, respectively).
In studies 309 and 310, the net benefit difference vs placebo was 12% (95% CI 4.28, 19.36) and 16%

Assessment report
EMA/178275/2019 Page 113/125
(95% CI 9.17, 23.43) for sotagliflozin 200 mg and 22% (95% CI 14.10, 29.64) and 17% (95% CI 10.06,
24.35) for sotagliflozin 400 mg. Again the main contributing component was the proportion of patients
achieving HbA1c <7.0% (36.9% and 33.3% in the 200 mg group, 46.9% and 33.8% in the 400 mg group
vs 22.8% and 15.1% placebo, study 309 and 310 respectively).

When the net benefit was analysed by its components (HbA1c < 7.0% without SH or DKA), the findings
were consistently in favour of sotagliflozin treatment. Generally higher proportions were observed with
the higher dose. All outcomes were statistically significant.

PPG was only evaluated in the CGM substudy which included 278 patients from studies 309 and 310. Both
doses of sotagliflozin resulted in a decrease in PPG. The treatment difference vs placebo was statistically
significant: -35 mg/dL (95%CI -58.1, -11.4) for the 200 mg dose and -41 mg/dL (95%CI -63.6, -18.3)
for the 400 mg dose. The findings were consistent with the data from the phase 2 program.

At Week 24, across the three studies, the LS mean difference from placebo in FPG was statistically
significant for all sotagliflozin groups, ranging from ­9.8 to ­25.7 mg/dL (p <0.001 for all groups with the
exception of sotagliflozin 200 mg in study 309 [p = 0.036]).

A large proportion of subjects (>70%) included in the studies were overweight. In all studies, the initial
insulin optimising/run-in period was associated with a weight increase in all treatment groups, ranging
from 0.3 kg to 0.95 kg with the largest increase observed in studies 309 and 310. At week 24, the
treatment difference versus placebo was ­2.35 (study 309) and ­1.98 kg (study 310) with sotagliflozin
200 mg; it was ­3.45 (study 309) and ­2.58 kg (study 310) with sotagliflozin 400 mg (all p <0.001). In
study 312, the treatment difference vs placebo was ­2.98 kg (p <0.001). The DEXA substudy showed that
the decrease in body weight was largely due to a loss of fat. This is in line with findings for other
SGLT2-inhibitors.

Change in bolus insulin dose was a secondary endpoint in all three studies. In all studies the bolus insulin
dose increased in the placebo treated groups (+3.9 to +6.6%). The change from baseline in the
sotagliflozin treated groups ranged from -2.6 to -10.4% with a LS mean difference from placebo ranging
from -5.7 to -16.4%. All comparisons vs placebo, except for the 200 mg group in study 309, were
statistically significant (p<0.001). The change in basal insulin dose and total insulin dose was also
assessed in all three studies. The pattern for the basal insulin dose was similar to that observed for the
bolus doses but the percent changes were smaller than for the bolus doses. In the placebo groups, a LS
mean increase in the total insulin dose at week 24 ranging from +0.6 to +2.9% was observed whereas the
LS mean decrease in total insulin dose observed in the sotagliflozin treated groups ranged from – 4.3 to
-8.9%.

Sotagliflozin treatment resulted in a modest placebo-adjusted decrease in SBP of -2 mmHg (95%CI -3.2,
-0.7) in the 200 mg group and -3.5 mmHg (95%CI -4.7, -2.3) in the 400 mg group, pooled data. There
was no apparent difference in the treatment effect in the overall population compared to patients with
SBP ≥130 mmHg whereas a slightly larger reduction was observed in patients with SBP ≥140 mmHg.

The aim of the CGM substudy was to investigate the effect of sotagliflozin on blood glucose variability. The
data showed an increase in the time spent in the desired blood glucose range (70-180 mg/dL) from
52.2% to 57.8% with sotagliflozin 200 mg and from 50.7% to 64.2% with sotagliflozin 400 mg, whereas
no relevant change was observed for placebo. The difference vs placebo was statistically significant for
the 200 mg group (5.3%, 95%CI 0.64, 10.06) and the 400 mg group (11.7%, 95%CI 7.14, 16.28). The
difference was explained by a decrease in the time spent in the hyperglycaemic blood glucose range;
whereas the time spent in the hypoglycaemic blood glucose range remained unchanged.

Patient related outcomes were assessed using the DTSQs and DDS2 scores. The DTSQs scores (range 0
to 36) indicated a higher satisfaction with sotagliflozin compared to placebo as reflected by a statistically
significant increase in the score (+1.7 and +2.5 for sotagliflozin 200 mg and 400 mg respectively, all p

Assessment report
EMA/178275/2019 Page 114/125
<0.001) as well as a higher proportion of patients reporting a change in score of 3 or more (54% and 52
% for sotagliflozin vs 28% for placebo). The DDS2 scores (range 0 to 6) indicated less distress with
sotagliflozin treatment compared to placebo with a treatment difference vs placebo ranging from -0.3 (p
= 0.025) to ­0.8 (p <0.001) across the sotagliflozin groups.

One-year exploratory efficacy data was provided from studies 309 and 310. At 52 weeks there was still a
modest but statistically significant treatment difference between placebo and sotagliflozin of
about -0.25% for the 200 mg dose and -0.3% for the 400 mg dose. Net benefit had decreased somewhat
at week 52, now ranging from 7.2 to 13.4% as compared to 12 to 22% at week 24. This was mainly due
to a lower proportion of patients achieving HbA1c <7.0%. The body weight slightly increased up to week
52 in the placebo treated groups, whereas body weight remained stable and reduced in all sotagliflozin
treated groups after week 24 and up to week 52. No apparent change in the effect on total fat mass was
observed. The insulin doses remained essentially unchanged after week 24. Patient reported outcomes
measuring diabetes distress remained in favour of sotagliflozin.

Pooled subgroup analyses from studies 309/310 were provided for the two doses, 200 mg and 400 mg
(ES1 pool) as well as for the ES2 pool (all patients treated with sotagliflozin 400 mg). The analyses
showed consistent findings for all demographic and baseline characteristics tested with regards to the
change from baseline in HbA1c, except for patients with eGFR <60 where the 95%CI included zero in all
three analyses. The number of patients were few, thus the 95%CI were wide. The subgroup analysis
indicates that the effect on HbA1c is attenuated in patients with eGFR <60 mL/min/1.73 m2 (-0.28% for
the 200 mg dose and -0.21% for the 400 mg dose) but the data is uncertain due to the low number of
subjects (47) included in the analysis. The SmPC recommends that treatment should not be initiated in
patients with eGFR <60 mL/min/1.73 m2, although treatment may be maintained until eGFR falls to 45
mL/min/1.73 m2.

Notably, no patients aged 75 years or above were included in the subgroup analyses.

No direct comparison between the doses investigated in studies 309 and 310 (200 mg and 400 mg) was
made; instead both doses were compared to placebo. Numerically there was a consistent treatment
difference between the two doses in favour of the 400 mg dose although the treatment difference with
regards to HbA1c was very small. Taking the individual response into account, a dose increase could be
beneficial in some patients, especially in overweight patients without increased ketosis and in need of
further intensification of treatment.

3.3. Uncertainties and limitations about favourable effects

The outcome of the three Phase 3 studies showed consistent results in spite of the differences in study
design between studies 309/310 and study 312. Thus there are no important uncertainties with regards
to the favourable effects in the overall population.

No specific clinical studies have been conducted in special populations. The number of older subjects
included in the studies was low. The SmPC includes recommendations not to initiate treatment in patients
above the age of 75 which is considered adequate and sufficient.

The clinical data in patients with impaired renal function is scarce. It is postulated that effect is maintained
also in renally impaired patients due to the combined effect of the SGLT1 and SGLT2 inhibition. The data
provided is however not sufficient to clarify how much of the effect is mediated by the SGLT1 inhibition.

Data from the DCCT trial has shown that a decrease in HbA1c is correlated with a decrease in mortality in
patients with T1DM. However the long-term effect of stabilised blood glucose levels is less well
documented, thus the clinical significance of the decrease in the glucose variability observed in the
studies is uncertain.

Assessment report
EMA/178275/2019 Page 115/125
3.4. Unfavourable effects

For the safety evaluation, results of the phase 2 and 3 studies from T1DM subjects were pooled.
The primary safety pool (SAF-1) includes two 52 weeks phase 3 studies conducted in T1DM subjects
treated in the same way, i.e. with insulin optimization and committee oversight of insulin dosing. The
broad safety pool (SAF-3) includes T1DM subjects from both phase 2 and 3 studies (the two 52 weeks
studies in SAF-1 and a world-wide 24 weeks phase 3 study of sotagliflozin 400 mg vs placebo in T1DM
subjects with insulin therapy that was not optimized). SAF-3 represents the largest pool of data from
T1DM subjects.

In total, 1,915 T1DM subjects (1,290 PY) were treated with sotagliflozin with 1,444 subjects (75%)
treated for at least 24 weeks and 668 subjects (64%) treated for at least 52 weeks.

The incidence of diabetic ketoacidosis was dose-dependently increased for sotagliflozin. In SAF-1, the
exposure-adjusted incidence rate (EAIR) was 3.12, 4.19 and 0.21 subjects per 100 PY (2.9%, 3.8% and
0.2%) for sotagliflozin 200 mg, sotagliflozin 400 mg and placebo. In SAF-3, the EAIR was 3.07 and 5.29
per 100 PY for sotagliflozin 200 mg and 400 mg, respectively, compared to 0.76 per 100 PY for placebo.
The incidence of DKA increased over time for sotagliflozin as compared to placebo in a clearly
dose-dependent manner. In SAF-1, the cumulative incidence of DKA at week 52 was approximately 2.8%
and 4.1% for sotagliflozin 200 mg and 400 mg sotagliflozin compared to 0.2% for placebo. The Kaplan
Meier plot of the cumulative incidence of adjudicated DKA in SAF-3 showed a more pronounced
differentiation between the 200 mg and 400 mg dose of sotagliflozin, with earlier TTO with 400 mg
compared to the SAF-1 plot. This may reflect the data from study 312 where no insulin optimization
occurred prior to randomization. Mean time to onset of DKA was shorter in sotagliflozin 400 mg (159
days) than in sotagliflozin 200 mg (208 days); although SD standard deviation was broad the difference
suggests a shorter TTO with sotagliflozin 400 mg compared to sotagliflozin 200 mg.

The incidence of documented hypoglycaemia was ≥98% in all treatment groups. The event rate was 81,
84 and 96 events per PY and the incidence of severe hypoglycaemia was 5.7%, 4.4% and 7.4%,
respectively, for sotagliflozin 200 mg, sotagliflozin 400 mg and placebo. In the Kaplan-Meier plot of the
cumulative incidence of severe hypoglycaemia, the sotagliflozin curves became less steep over time and
provided a separation from placebo after about 24 weeks of treatment.

Events of volume depletion were reported more frequently with sotagliflozin (200 mg: 2.7%; 400 mg:
1.1%) than with placebo (1.0%). In SAF-3, subjects ≥65 years of age (n=209) had a higher incidence of
volume depletion for sotagliflozin 200 mg and 400 mg relative to placebo and subjects with eGFR <60
mL/min/1.73 m2 (n=146) had a higher incidence of volume depletion for sotagliflozin 400 mg versus
placebo.

There was an increased risk of genital infections in sotagliflozin-treated subjects relative placebo. Both
female (15% and 17% vs 4.7%) and male (3.0% and 6.3% vs 1.1%) genital infections were increased
compared to placebo. Most of the events were mild to moderate in intensity.

The incidence of urinary tract infections in female subjects was increased and similar in subjects treated
with sotagliflozin 200 mg (12%) and placebo (11%) compared to sotagliflozin 400 mg (7.0%); however,
the incidence of UTI in male subjects was dose-dependently increased in the sotagliflozin 200 mg and 400
mg group (2.3% and 4.0%, respectively) as compared to placebo (1.8%).

The incidence of diarrhoea was increased in the sotagliflozin 200 mg and 400 mg group (6.5% and 8.8%,
respectively) as compared to placebo (5.1%). Most of the events were mild or moderate in intensity.

Treatment with sotagliflozin was associated with a decrease in eGFR in the sotagliflozin 200 mg and 400
mg groups (-4.0% and -4.3%) compared to placebo (-1.3%) that that did return towards baseline at
week 52. Also in subjects with moderate and mild renal impairment, the eGFR returned towards baseline

Assessment report
EMA/178275/2019 Page 116/125
at week 52, however; subjects with eGFR <60 mL/min/1.73 m2 made up for a small number in SAF-1
(n=71), wherefore difficult to draw any firm conclusions in this subgroup. The incidence of renal-related
events was similar across the sotagliflozin groups (1.5% and 1.3%) and placebo (1.5%).

The incidence of fractures through 52 weeks of treatment was similar across the groups: 2.9%, 1.9% and
3.4% for sotagliflozin 200 mg, sotagliflozin 400 mg and placebo. A statistically significant reduction in
‘lumbar spine’ bone mineral density was observed at week 52 for sotagliflozin 200 mg [LS Mean difference
from placebo -1.27% (-2.51%, -0.03%), p=0.044] and sotagliflozin 400 mg [LS Mean difference from
placebo -2.53% (-3.79%, -1.28%), p<0.001] and in ‘total hip’ BMD for sotagliflozin 400 mg [LS Mean
difference from placebo -1.18% (-2.07%, -1.28%), p=0.009]. There was a dose-dependent increase
from baseline to week 52 in the bone resorption marker, CTX, for sotagliflozin 200 mg and 400 mg (18%
and 32%) relative to placebo (1.2%) and a non-dose-dependent increase in the bone formation marker,
P1NP (8.7% and 3.4% vs. -4.9% for sotagliflozin 200 mg and 400 mg vs. placebo).

Subgroups

In subgroup analyses in the SAF-3 pool, there was rather limited data for subjects ≥65 years of age and
subjects with eGFR <60 mL/min/1.73 m2 to draw any firm conclusions.

In subjects ≥65 years of age (n=209), there was an increased risk of events related to volume depletion
for sotagliflozin (102 and 38 EAIR per 1000 PY for sotagliflozin 200 mg and 400 mg, respectively)
compared to placebo (zero events).

In subjects with eGFR 60 to <90mL/min/1.73 m2 (n=1,492), the incidence of events of volume depletion
was higher for sotagliflozin (28 and 24 subjects per 1000 PY for sotagliflozin 200 mg and 400 mg)
compared to placebo (8.3 subjects per 1000 PY). In subjects with eGFR <60 mL/min/1.73 m2 (n=146),
the incidence of volume depletion was 111 subjects per 1000 PY for sotagliflozin 400 mg and 0 events for
sotagliflozin 200 mg and placebo, respectively.. Sotagliflozin should not be initiated in subjects with
moderate renal impairment (eGFR <60 mL/min), as advised in the SmPC.

3.5. Uncertainties and limitations about unfavourable effects

Overall, there is rather limited data from the clinical studies in subjects ≥65 years (n=265) and very
limited data in subjects ≥75 years (n=18). Accordingly, the SmPC advises that initiation of sotagliflozin
treatment in patients above 75 years of age is not recommended.

The major safety concern of sotagliflozin is the increased and dose-dependent risk of DKA. The incidence
of DKA increased over time for sotagliflozin as compared to placebo in a clearly dose-dependent manner.
In SAF-1, the cumulative incidence of DKA at week 52 was approximately 2.8% and 4.1% for sotagliflozin
200 mg and 400 mg sotagliflozin compared to 0.2% for placebo.

Possible risk factors/contributing factors associated with an event of a positively adjudicated DKA were
analysed. However, the analyses of potential risk factors of DKA were associated with uncertainties due to
low number of events. However, the subgroup analyses indicated that subjects with previous DKA,
subjects with high BHB levels at baseline or greater increase in BHB levels, subjects who had their insulin
dose decreased by more than 20% and subjects who were CSII-users, were at higher risk to develop DKA.
This has been reflected in the proposed product information. In 19% (7/35) of the cases, no cause (no
contributing factor/ risk factor) for the DKA event could be identified.

The applicant has proposed risk minimisation measures in order to mitigate the risk of DKA which includes
education of health care professionals and patients about self-monitoring of ketones, situations at-risk to
allow for early diagnosis of ketonaemia or ketonuria and how to manage potential ketosis and
maintenance of optimal insulin therapy. Detailed recommendations are provided in the SmPC and in the

Assessment report
EMA/178275/2019 Page 117/125
proposed educational materials. These recommendations are considered adequate in order to mitigate
the risk, based on the knowledge on DKA in general and on the experience from the clinical trials.

3.6. Effects Table

Table 29 Effects Table for Zynquista in the treatment of T1DM.

Effect Short Unit Treatment Control Uncertainties/ Refere
Description Strength of nces
evidence

Favourable Effects

HbA1c Change from % 200 mg Placebo Difference (95%CI)

baseline at -0.43 -0.07 -0.36 (-0.45, -0.27) Study
week 24 -0.39 -0.02 -0.37 (-0.48, -0.25) 309 and
310
400 mg
-0.48 -0.41 (-0.50, -0.32)
-0.37 -0.35 (-0.47, -0.24)

400 mg
-0.79 -0.33 -0.46 (-0.54, -0.38) Study
312
Net Proportion of % 200 mg Placebo Difference (95%CI)
benefit patients with 33.5 21.6 11.8 (4.28, 19.36) Study
A1C <7.0% 31.4 15.1 16.3 (9.17, 23.43) 309 and
and no SH and 310
no DKA at 400 mg
week 24 43.5 21.9 (14.10, 29.64)
32.3 17.2 (10.06, 24.35)

400 mg
28.6 15.2 13.4 (9.12, 17.67) Study
312
Bolus Change from % 200 mg Placebo Difference (95%CI)
insulin baseline at -1.8 +3.9 -5.7 (-12.82, 1.42) Study
dose week 24 -7.0 +5.9 -12.9 (-20.50, -5.38) 309 and
310
400 mg
-8.8 -12.7 (-19.79, -5.55)
-10.5 -16.4 (-23.90, -8.83)

400 mg
-5.7 +6.6 -12.3 (-18.17, -6.48) Study
312
Body Change from kg 200 mg Placebo Difference (95%CI)
weight baseline at -1.57 +0.78 -2.35 (-2.85, -1.85) Study
week 24 -1.88 +0.17 -1.98 (-2.53, -1.44) 309 and
310
400 mg
-2.67 -3.45 (-3.95, -2.94)
-2.47 -2.58 (-3.12, -2.04)

400 mg
-2.21 +0.77 -2.98 (-3.31, -2.66) Study
312

Assessment report
EMA/178275/2019 Page 118/125
Effect Short Unit Treatment Control Uncertainties/ Refere
Description Strength of nces
evidence
Glucose Percent time % 200 mg Placebo Difference (95%CI) CGM
variability spent inside 57.8 51.6 5.3 (0.64, 10.06) sub-stu
target range P = 0.026 dy
(70-180 mg/d 400 mg
L) 64.2 11.7 (7.14, 16.28)
p <0.001
Systolic Changes from mmHg 200 mg Placebo Difference (95%CI) Study
blood baseline at -2.6 +0.6 -3.5 (-5.2, -1.8) 309 and
pressure week 12 -2.2 -1.8 -0.4 (-2.2, 1.3) 310

400 mg
-3.3 -4.2 (-5.9, -2.4)
-4.8 -2.8 (-4.6, -1.1)

400 mg Study
-2.21 +0.5 -3.8 (-4.9, -2.7) 312

DTSQs, Change in % 200 mg Placebo Study

responder score of 3 or 54, 41 28, 29 p <0.005, all 309 and
more comparisons 310
400 mg
52, 43

Unfavourable Effects

Diabetic Number of n (%) 200 mg: 1 (0.2%) SAF-1

keto- subjects with 15 (2.9%)
acidosis definite DKA
400 mg:
20 (3.8%)

Diabetic Exposure-adj EAIR 200 mg: 0.21 Rate of DKA in T1DM SAF-1
keto- usted (n/ 3.12 subjects per reported in the
acidosis incidence rate 100 subjects per 100 PY literature to be 2
(subjects per PY) 100 PY subjects per 100 PY
100 (or 3%)
patient-years) 400 mg:
4.19
subjects per
100 PY

Diabetic Number of n 200 mg: 6 SAF-3

keto- subjects with 15
acidosis definite DKA
400 mg:
42

Diabetic Exposure-adj EAIR 200 mg: 0.76 Rate of DKA in T1DM SAF-3
keto- usted (n/ 3.07 subjects per reported in the
acidosis incidence rate 100 subjects per 100 PY literature to be 2
(subjects per PY) 100 PY subjects per 100 PY
100 (or 3%)
patient-years) 400 mg:
5.29
subjects per
100 PY

Assessment report
EMA/178275/2019 Page 119/125
Effect Short Unit Treatment Control Uncertainties/ Refere
Description Strength of nces
evidence
Severe Subjects with n (%) 200 mg: 39 (7.4%) SAF-1
hypo- at least 1 30 (5.7%)
glycaemia event
400 mg:
23 (4.4%)

Volume Incidence of % 200 mg: 1.0% SAF-1

depletion volume 2.7%
depletion
400 mg:
1.1%
Urinary Incidence of % 200 mg: 6.1% SAF-1
tract urinary tract 7.1%
infections infections
400 mg:
5.5%
Genital Incidence of % Female Female SAF-1
mycotic genital genital genital
infections mycotic infections infections
infections 200 mg: 4.7%
15%

400 mg:
17%

Male genital Male genital

infections infections
200 mg: 1.1%
3.0%

400 mg:
6.3%

3.7. Benefit-risk assessment and discussion

3.7.1. Importance of favourable and unfavourable effects

T1DM is characterised by insulin deficiency due to destruction of the insulin-producing cells. Insulin
treatment aims at normalising blood glucose levels in order to avoid acute symptoms of hyperglycaemia
and to minimise the risk of long-term microvascular and macrovascular complications. Optimal treatment
requires that the patient monitors blood glucose levels and makes insulin dose adjustments on a daily
basis. Thus, the management of T1DM has a large impact on the patient’s daily life. Currently only insulin
is approved for the treatment of T1DM in the EU. In spite of improvements in insulins, methods of
administration and monitoring of blood glucose, normalisation of glucose levels is difficult, and the
treatment is associated with hypo- and hyperglycaemia as well as weight increase. Thus, there is a need
for new therapies as an adjunct to insulin therapy, in order to alleviate the negative effects of insulin
treatment in order to reach treatment targets. In addition, treatment of patients with T1DM carries an
inherent risk of DKA since interruption of treatment or an excessive increase in insulin need will result in
the development of DKA.

The data provided with this application show that sotagliflozin, when added to insulin, consistently
reduced HbA1c by about 0.4% from a relatively low baseline HbA1c of 7.6-7.7%. This was achieved with
a reduction of the risk of severe hypoglycaemia compared to placebo. Sotagliflozin was also shown to
reduce glucose variability as indicated by a significant increase in time spent in the desired blood glucose

Assessment report
EMA/178275/2019 Page 120/125
range. Although the clinical significance of time spent in range is not strongly established, the applicant
refers to a recent analysis of DCCT data which shows that there is an association between time in range
and the development of microvascular complications. Sotagliflozin also had consistent and clinically
relevant effects on the CV risk factors body weight and SBP. In addition, patients treated with sotagliflozin
reported a significant improvement in treatment satisfaction and distress due to the disease. Notably,
these patient-related outcomes were observed in spite of the increased ketone monitoring used in the
trials in order to reduce the risk of DKA. Thus, the effect of sotagliflozin is considered to be of clear clinical
relevance.

The effect in different subgroups was comparable to that of the total population. The unmet need in
patients with overweight/obesity is however very different compared to patients with normal weight
considering that the alternative treatment, i.e. increase of insulin dose, will lead to additional weight gain
which subsequently may increase the risk of cardiovascular disease. Therefore, the expected benefit in
clinical practice is higher in overweight/obese patients compared to the total T1DM population.

There were no apparent differences between the two doses investigated with regards to the effect on
glycaemic control; however, a more pronounced effect on body weight was observed with the higher
dose, especially in obese subjects. There is a trend towards a better effect on time in range and blood
pressure and the reduction of severe hypoglycaemias was more pronounced with the 400 mg dose.
Taking the individual response into account, a dose increase could be beneficial in some patients and the
higher dose is therefore considered justified.

The safety profile of sotagliflozin is in general as expected and consistent with other SGLT2 inhibitors
regarding the increased risk of volume depletion, genital infections and UTI. Inhibition of SGLT1 by
sotagliflozin delays and reduces glucose absorption in the proximal intestine, which can lead to diarrhoea.

There are limitations in both safety and efficacy data in elderly patients (>75 years of age) and patients
with moderate renal impairment (eGFR <60 mL/min/1.73 m2). Initiation of sotagliflozin treatment in
subjects above 75 years of age is not recommended. Sotagliflozin should not be initiated in subjects with
moderate renal impairment (eGFR <60 mL/min).

In 2015, an Art 20 referral procedure on SGLT2-inhibitors and DKA (EMA/PRAC/50218/2016) was

initiated due to an increased reporting of DKA in T2DM patients. The data available at the time of the
procedure was not sufficient to conclude whether SGLT2-inhibitors increased the risk of DKA. However,
based on the knowledge about the pharmacodynamic effect of SGLT2-inhibition, it is plausible that
treatment with SGLT2-inhibitors could promote DKA development.

The data presented with this application indeed show that, in spite of the precautionary measures taken,
there was a considerable increase in the risk of DKA compared to placebo in T1DM patients. This lends
support to a direct promoting effect of sotagliflozin on the development of DKA. The increased risk is of
concern, considering that DKA is a condition which is potentially life-threatening that require
hospitalisation.

Possible risk factors/contributing factors associated with an event of a positively adjudicated DKA was
analysed. However, the analyses of potential risk factors of DKA are associated with uncertainties due to
low number of events. However, the subgroup analyses indicated that subjects with previous DKA,
subjects with high BHB levels at baseline or greater increase in BHB levels, subjects who had their insulin
dose decreased with more than 20% and subjects who were CSII-users, were at higher risk to develop
DKA. This has been reflected in the proposed product information.

The applicant has proposed risk minimisation measures in order to mitigate the risk of DKA which include
education of health care professionals and patients about self-monitoring of ketones, situations at-risk to
allow for early diagnosis of ketonaemia or ketonuria and how to manage potential ketosis and
maintenance of optimal insulin therapy. Detailed recommendations are provided in the SmPC and in the

Assessment report
EMA/178275/2019 Page 121/125
proposed educational materials. These recommendations are considered adequate in order to mitigate
the risk, based on the knowledge on DKA in general and on the experience from the clinical trials.

However, as the risk of DKA may not be totally abolished, the benefit risk balance in the initially proposed
target population, i.e. all patients with T1DM, is considered negative.

However, overweight patients (BMI ≥27 kg/m2) with T1DM and inadequate glycaemic control is a
population with an unmet medical need since currently the only treatment option is intensified insulin
treatment, known to result in further increases in body weight. In addition to the blood glucose lowering
effect, the weight reduction and the reduction of SBP are of greatest importance in patients with high BMI.
Furthermore, patients with higher BMI have a higher insulin need and as indicated by the data, higher
insulin doses decrease the risk of DKA. Therefore, the benefits are considered to outweigh the risks in
patients with BMI ≥27 kg/m2 who have failed to achieve desired glycaemic control despite optimal insulin
therapy.

3.7.2. Balance of benefits and risks

The target population is restricted to patients with a BMI ≥27 kg/m2. In this population the benefit risk
balance is positive.

3.7.3. Additional considerations on the benefit-risk balance

An AHEG meeting was held 21 November 2018, see section 2.6.1. (Discussion on clinical safety).

A PASS study (category 1) is planned. This is a non-interventional study to evaluate the risk of DKA of
sotagliflozin in patients with type 1 diabetes mellitus (T1DM) in a real-world setting. The objective is to
evaluate the incidence of DKA with sotagliflozin treated patients in real life as compared to non-SGLTi
treated patients to determine effectiveness of additional risk-minimisation measures for DKA in the post
approval setting.

3.8. Conclusions

The overall B/R of Zynquista is positive.

4. Recommendations

Outcome

Based on the CHMP review of data on quality, safety and efficacy, the CHMP considers by consensus that
the benefit-risk balance of Zynquista is favourable in the following indication:

Zynquista is indicated as an adjunct to insulin therapy to improve glycaemic control in adults with type 1
diabetes mellitus with a Body Mass Index (BMI) ≥ 27 kg/m2, who have failed to achieve adequate
glycaemic control despite optimal insulin therapy.

The CHMP therefore recommends the granting of the marketing authorisation subject to the following
conditions:

Assessment report
EMA/178275/2019 Page 122/125
Conditions or restrictions regarding supply and use

Medicinal product subject to restricted medical prescription .

Other conditions and requirements of the marketing authorisation

Periodic Safety Update Reports

The requirements for submission of periodic safety update reports for this medicinal product are set out
in the list of Union reference dates (EURD list) provided for under Article 107c(7) of Directive 2001/83/EC
and any subsequent updates published on the European medicines web-portal.

The marketing authorisation holder shall submit the first periodic safety update report for this product
within 6 months following authorisation.

Conditions or restrictions with regard to the safe and effective use of the
medicinal product

Risk Management Plan (RMP)

The MAH shall perform the required pharmacovigilance activities and interventions detailed in the agreed
RMP presented in Module 1.8.2 of the marketing authorisation and any agreed subsequent updates of the
RMP.

An updated RMP should be submitted:

• At the request of the European Medicines Agency;

• Whenever the risk management system is modified, especially as the result of new information
being received that may lead to a significant change to the benefit/risk profile or as the result of
an important (pharmacovigilance or risk minimisation) milestone being reached.

Additional risk minimisation measures

Prior to launch of Zynquista (sotagliflozin), as an adjunct to insulin therapy to improve glycaemic control
in adults with type 1 diabetes mellitus with a Body Mass Index (BMI) ≥ 27 kg/m2, who have failed to
achieve adequate glycaemic control despite optimal insulin therapy, in each Member State, the Marketing
Authorisation Holder (MAH) must agree the content and format of educational materials for sotagliflozin,
including communication media, distribution modalities, and any other aspects of the programme, with
the National Competent Authority.

The educational materials are aimed at providing guidance on how to manage risk of diabetic ketoacidosis
(DKA) in patients with type 1 diabetes.

The MAH shall ensure that in each Member State where sotagliflozin is marketed, all healthcare
professionals and patients/carers who are expected to prescribe, dispense or use the product have access
to:

• Guide for Health Care Professionals including a prescriber’s checklist

• Patient’s/Carer’s Guide
• Patient Alert Card

The guide for healthcare professionals including the prescriber’s checklist should contain the following key
elements:

Assessment report
EMA/178275/2019 Page 123/125
• Sotagliflozin is not a substitute for insulin (and does not alter insulin-sensitivity).
• The risk of DKA is increased with sotagliflozin treatment.
• If treated with sotagliflozin, glucose levels will not adequately reflect insulin needs, and DKA may
occur in patients treated with sotagliflozin even if blood glucose levels are below 14 mmol/l (250
mg/dl). Therefore, glucose monitoring must be supplemented by ketone monitoring.
• Patients with euglycaemic DKA may need glucose in addition to standard of care for DKA and
sotagliflozin should be discontinued if DKA occurs.
• Guidance to the physician for assessing whether the patient is eligible for sotagliflozin prescription,
e.g. patient selection criteria including adherence to insulin treatment and insulin thresholds,
patient’s beta-hydroxybutyrate (BHB) < 0.6 mmol/L or urine ketones < 1+, BMI ≥ 27 kg/m2, absence
of DKA risk factors.
• Guidance to the physician for assessing whether the patient is prepared and engaged to perform
self-ketone testing before and during therapy.
• Summary of the recommendations for patients, particularly regarding blood ketone measurement
and managing sick days.
• For pump users: restrict sotagliflozin prescription to patients experienced in pump use, common
trouble-shooting strategies when interruptions of insulin delivery via pump occur in case of pump
failure.
• Counsel the patient and evaluate their adherence to ketone monitoring while establishing their
baseline ketone level 1 to 2 weeks before treatment initiation and ensure the patient
o Has received education/training in ketone testing and interpreting/acting upon test results
o Is willing/able to perform ketone testing as prescribed
o Is adequately informed about managing sick days
• Ensure the patient is on optimal insulin therapy prior to initiation of sotagliflozin treatment.
• Sotagliflozin treatment should be temporarily stopped before surgical procedures or in case of
hospitalisation for acute serious illness.
• If addition of sotagliflozin leads to marked reduction of insulin need, discontinuation of sotagliflozin
should be considered to avoid high risk of DKA.

The patient’s/carer’s guide should contain the following key elements:

• Sotagliflozin is not a substitute for insulin.

• DKA may occur in patients treated with sotagliflozin even if blood glucose levels are below 14 mmol/l
(250 mg/dl), i.e. an explanation of the concept of euglycaemic DKA.
• Signs/symptoms of DKA - if not adequately managed DKA can be severe and fatal.
• How to measure ketones, how to interpret the results and what to do in case of hyperketonaemia/DKA
(contact HCP immediately if BHB > 0.6 mmol/L with symptoms or if BHB > 1.5 mmol/L with or without
symptoms).
• Insulin dose reduction during treatment should only be done when needed to prevent hypoglycaemia
and should be done cautiously to avoid ketosis and DKA.
• Do not start caloric restriction or carbohydrate restriction while treated with sotagliflozin.

The patient alert card should contain the following key elements:

• The patient alert card should be presented to any HCP consulted.

• DKA may occur in patients treated with sotagliflozin even if blood glucose levels are below 14 mmol/l
(250 mg/dl).
• Signs/symptoms of DKA.
• Patients with euglycaemic DKA should receive glucose, insulin and fluids for DKA, sotagliflozin should
be discontinued.
• Sotagliflozin should be temporarily stopped before surgical procedures or hospitalisation for acute
serious illness.
• Contact details of the sotagliflozin prescriber’ and ‘Name of patient’.

Obligation to conduct post-authorisation measures:

The MAH shall complete, within the stated timeframe, the below measures:

Description Due date

Assessment report
EMA/178275/2019 Page 124/125
 Description Due date

Non-interventional PASS: In order to estimate the incidence of DKA in T1DM 31/12/2024

sotagliflozin treated patients to assess the effectiveness of the risk minimisation
measures implemented in Europe, the MAH should conduct and submit the results
from an observational cohort study using existing data sources in European countries
where sotagliflozin will be launched for T1DM.

Conditions or restrictions with regard to the safe and effective use of the
medicinal product to be implemented by the Member States

Not applicable.

These conditions fully reflect the advice received from the PRAC.

New Active Substance Status

Based on the CHMP review of the available data, the CHMP considers that sotagliflozin is a new active
substance as it is not a constituent of a medicinal product previously authorised within the European
Union.

Assessment report
EMA/178275/2019 Page 125/125