CENTER FOR DRUG EVALUATION AND

RESEARCH

APPLICATION NUMBER:

204629Orig1s000

SUMMARY REVIEW
 Division Director Review

On June 3rd, 2014 Boehringer Ingelheim Pharmaceuticals, Inc. submitted a Class 1

resubmission of the new drug application for Jardiance under section 505(b)(1) of the Federal
Food, Drug and Cosmetic Act. The resubmission is a complete response to the Complete
Response Letter issued by the Agency on March 4th 2014. The application received a
complete response due to product quality issues. In brief, FDA inspections revealed
violations of current good manufacturing practices (cGMP) at the empagliflozin drug
substance and product manufacturing site. Drug substances (i.e., APIs) and finished
pharmaceutical products manufactured at Boehringer-Ingelheim Pharma GmbH & Co. KG in
Ingelheim am Rhein, Germany were deemed adulterated within the meaning of Section
501(a)(2)(B) of the Federal Food, Drug and Cosmetic Act, 21 USC 351(s)(2)(B) as a result of
these violations. Boehringer Ingelheim Pharma GmbH & Co. KG was issued a Warning Letter
on May 6th 2013. The firm took corrective action and FDA re-inspected this facility during the
period of February 24th-March 7th 2014. The FDA agreed that corrective actions addressed
the violations listed in the May 6th 2013 Warning Letter and re-classified the manufacturing
site as acceptable on May 28th 2014. Refer to Dr. Leginus’ memorandum for details. This
resolves the deficiency listed in the Complete Response Letter.

As part of the re-submission, the applicant also provides an update to the major safety
information from the original NDA with data derived from one recently completed trial and
extensions of parent trials previously reviewed. Dr. Chong has reviewed the newly submitted
safety information in detail in his CDTL memorandum and concludes that the additional
safety data does not change the favorable benefit risk profile established with data from the
original NDA. I have reviewed his memorandum and I am in agreement with his conclusion.

I recommend approval of the application pending final agreement on labeling. The applicant
has addressed the manufacturing issues adequately and the additional safety data do not
change the favorable benefit-risk profile established with data in the original submission. For
a full discussion of benefit risk with Jardiance refer to my previous memorandum dated
March 4th 2014. The following post-marketing studies will be required under Sections
505(o)(3) or Sections 505B(a) of the Federal Food, Drug and Cosmetic Act. Refer to the PMR
development template for details.

 Completion of an ongoing cardiovascular outcomes trial to assess the cardiovascular

risk associated with Jardiance use and to monitor for the following events of interest:
renal failure, specific malignancies [lung, bladder, breast and melanoma], drug
induced liver injury, complicated genital and urinary tract infections, hypotension,
fractures, and serious hypersensitivity reactions.
 A pediatric pharmacokinetic/pharmacodynamic study, and a pediatric safety and
efficacy study. As part of the safety and efficacy study, the effect on bone health and
development will be evaluated.
 A nonclinical juvenile toxicity study with a particular focus on renal development,
bone development, and growth.

Page 2 of 2

Reference ID: 3602846

 ---------------------------------------------------------------------------------------------------------
This is a representation of an electronic record that was signed
electronically and this page is the manifestation of the electronic
signature.
---------------------------------------------------------------------------------------------------------
/s/
----------------------------------------------------
JEAN-MARC P GUETTIER
07/31/2014

Reference ID: 3602846

 Division Director Review

Division Director Memorandum

1. Introduction

On March 5, 2013 Boehringer Ingelheim submitted a 505(b)(1) new drug application (NDA)
for Jardiance. The applicant is seeking to indicate Jardiance as an adjunct to diet and exercise
to improve glycemic control in adults with type 2 diabetes mellitus. Jardiance is a tablet
containing either 10 or 25 mg of empagliflozin [i.e., a selective sodium glucose co-transporter
2 (SGLT-2) inhibitor]. If approved, Jardiance will be the third SGLT-2 inhibitor indicated for
use in the management of patients with type 2 diabetes mellitus in the United States.

This document serves as the division director’s memorandum for the application.

2. Background

The sodium glucose cotransporter 2 (SGLT-2) is a low-affinity, high-capacity, active sodium

glucose symporter expressed at the apical membrane of epithelial cells lining the proximal
renal tubule. SLGT-2 is responsible for reabsorbing the majority of the glucose filtered at the
glomerulus. Inhibition of SGLT-2 by empagliflozin decreases glucose reabsorption and
promotes urinary glucose excretion. Urinary glucose loss is the primary mechanism by which
empagliflozin lowers blood glucose.

The glucose lowering effect of agents that work by inhibiting SGLT-2 wanes as renal function
deteriorates. Invokana and Farxiga, the two approved products in the class, did not provide
clinically meaningful glucose lowering in patients with estimated glomerular filtration rate
below 45 and 60 mL/min/1.73 m2 respectively.

Inhibition of glucose reabsorption induces an osmotic diuresis resulting in polyuria and

intravascular volume contraction. Adverse reactions associated with volume contraction
identified in clinical trial datasets for the two approved products include hypotension and
impairment in renal function. Advanced age, impairment in renal function and use of specific
diuretics at baseline were identified as risk factors for volume contraction related adverse
reactions in these programs.

Promotion of glucosuria, as a result of SGLT-2 inhibition, increases the risk for male and
female genital mycotic infections. These class-related adverse reactions were observed with
use of both Invokana and Farxiga products.

The development program for empagliflozin was discussed with the Division of Metabolism
and Endocrinology Products (DMEP). Records of these interactions can be found in an End of
Phase II Advice letter issued December 8, 2009, End of Phase II Meeting Minutes issued June

Page 2 of 15

Reference ID: 3464120

 Division Director Review

3, 2010, Written Responses to proposed analyses issued May 22, 2012 and pre-NDA Meeting
Minutes issued December 17, 2012. The development program included a prospective
proposal to assess cardiovascular risk associated with empagliflozin use to satisfy the
requirements stipulated in the 2008 FDA Guidance for Industry entitled: Diabetes Mellitus-
Evaluating Cardiovascular Risk in New Antidiabetic Therapies to Treat Type 2 Diabetes. The
acceptability of the proposed plan was reviewed by DMEP as reflected in an advice letter
issued on April 9th 2012.

3. CMC/Device
I concur with the conclusions reached by Dr. Leginus regarding the acceptability of the
proposed manufacturing processes for the drug substance and drug product. Stability testing
supports an expiry of 36 months when the product is stored at room temperature.

The drug substance and finished products will be manufactured by Boehringer-Ingelheim

Pharma GmbH & Co. KG., in Ingelheim am Rhein, Germany. Manufacturing site inspections
performed in November 2012 identified significant violations of current good manufacturing
practices (cGMP) for the manufacture of active pharmaceutical ingredients (API) and finished
pharmaceutical products at that site. Drug substances (i.e., APIs) and finished
pharmaceutical products manufactured at Boehringer-Ingelheim Pharma GmbH & Co. KG.
were deemed adulterated within the meaning of Section 501(a)(2)(B) of the Federal Food,
Drug and Cosmetic Act, 21 USC 351(s)(2)(B) as a result of these violations. Boehringer
Ingelheim Pharma GmbH & Co. KG. was issued a Warning Letter on May 6th 2013 detailing
these violations.

Since issuance of the Warning letter, the firm has taken corrective actions and notified the
Division that they were ready for re-inspection on January 8th 2014. The re-inspection will be
extensive and will include a full cGMP inspection and multiple pre-approval inspections. It is
scheduled to take place between February 24th and March 7th 2014. The re-inspection, and
preparation of the Establishment Inspection Report will not be completed until after the
Prescription Drug User Fee Act goal date of March 5th 2014. Compliance will make a withhold
approval recommendation until they can establish that drug substances and finished
products manufactured at Boehringer Ingelheim Pharma GmbH & Co. KG. are no longer
deemed adulterated within the meaning of Section 501(a)(2)(B) of the Federal Food, Drug
and Cosmetic Act, 21 USC 351(s)(2)(B)s.

4. Nonclinical Pharmacology/Toxicology
I concur with the conclusions reached by Drs. Summan and Bourcier that there are no
outstanding pharm/tox issues that preclude approval.

Page 3 of 15

Reference ID: 3464120

 Division Director Review

5. Clinical Pharmacology/Biopharmaceutics
I concur with the conclusions reached by the clinical pharmacology/biopharmaceutics review
team that there are no outstanding clinical pharmacology issues that preclude approval.

6. Clinical Microbiology
Not applicable.

7. Clinical/Statistical-Efficacy
To support the indication of improved glycemic control, the long term glucose lowering effect
of empagliflozin was evaluated in 4 pivotal phase 3 clinical trials. These trials were multi-
center, multi-national, randomized, double blind and placebo-controlled. The variable used
in the primary efficacy assessment was the difference in the change in hemoglobin A1c (i.e.,
HbA1c) from baseline to trial end between empagliflozin-treated subjects and placebo-
treated subjects. Efficacy was assessed at 24-weeks in each of these trials. Inclusion and
exclusion criteria were similar between trials and these are reviewed in Drs. Chong and Liu’s
reviews. The design features, primary endpoint and timing of the efficacy assessment in the
empagliflozin phase III pivotal trials conform with the Guidance for Industry entitled
“Diabetes Mellitus: Developing Drugs and Therapeutic Biologics for Treatment and
Prevention” and are appropriate to support an indication of improved glycemic control.

The efficacy of empagliflozin was assessed in various, relevant, clinical use settings that
included;

 Empagliflozin used as monotherapy in drug naïve adult subjects with type 2 DM not
optimally controlled1 on diet and exercise alone.

o Trial 1245.20 – Efficacy assessment at 24 weeks

o Compared empagliflozin (10 and 25 mg) to placebo (primary objective) and
sitagliptin (100 mg dose and secondary objective)
o Eligible2 completers were asked to re-consent to enter trial 1245.31 a double-blind
controlled extension of at least 52 weeks beyond the primary endpoint3

 Empagliflozin used as add-on therapy to background metformin in adult subjects with

type 2 DM not optimally controlled1 on ≥ 1500 mg/day of metformin

o Trial 1245.23 met – Efficacy assessment at 24 weeks

1
HbA1c >7.0% and <10.0%
2
As an example, completers with ALT or AST > 3X the ULN, with eGFR < 30 mL/min or < 60 mL/min (China), or
who were known to have hypersensitivity to product (Ireland) at trial end were excluded.
3 th
database cutoff for NDA submission May 29 2012

Page 4 of 15

Reference ID: 3464120

 Division Director Review

o Compared empagliflozin (10 and 25 mg) to placebo

o Eligible completers were asked to re-consent to enter trial 1245.31 a double-blind
controlled extension of at least 52 weeks beyond the primary endpoint3

 Empagliflozin used as add-on therapy to background metformin and sulfonylurea in

adult subjects with type 2 DM not optimally controlled1 on ≥ 1500 mg/day of
metformin and at least 4 mg of glimepiride.

o Trial 1245.23 met + SU – Efficacy assessment at 24 weeks

o Compared empagliflozin (10 and 25 mg) to placebo
o Eligible completers were asked to re-consent to enter trial 1245.31 a double-blind
controlled extension of at least 52 weeks beyond the primary endpoint3

 Empagliflozin used as add-on therapy to background pioglitazone or piolitazone and

metformin in adult subjects with type 2 DM not optimally controlled1 on ≥ 30 mg/day
of pioglitazone alone or in combination with metformin.

o Trial 1245.19 - Efficacy assessment at 24 weeks

o Compared empagliflozin (10 and 25 mg) to placebo
o Eligible completers were asked to re-consent to enter trial 1245.31 a double-blind
controlled extension of at least 52 weeks beyond the primary endpoint3

The primary efficacy results for these four pivotal trials, reproduced from Table 3 in Dr. Liu’s
statistical review are shown below. Improvement in glycemic control was demonstrated for
both doses across all four trials. The observed placebo adjusted least square mean reduction
in HbA1c across doses and trials at 24-weeks ranged from -0.48% to -0.83%. In three of the
four trials the 25 mg dose provided numerically greater HbA1c reduction than the 10 mg dose
though the magnitude of the added reduction gained by more than doubling the dose was
small. Dr. Liu performed a number of sensitivity analyses to test the robustness of the
primary analysis model and found the results to be robust.

Page 5 of 15

Reference ID: 3464120

 Division Director Review

Table 1: Summary of Primary Efficacy Endpoints in the 4 Pivotal Trials (Source Table 3 in Dr. Liu's review)
Number Baseline Change Froom Adjusted
Trial/ of subjects HbA1c Baseline at Mean 97.5% CI P-value
Treatment Group Mean (SE) Week 24 Difference
1245.20 (monotherapy)
Placebo 228 7.91 (0.05) 0.06 (0.05) --- --- ---
Empagliflozin 10 mg 224 7.87 (0.06) -0.66 (0.05) -0.72 (-0.89, -0.56) <0.0001
Empagliflozin 25 mg 224 7.86 (-.06) -0.77 (0.06) -0.83 (-0.99, -0.68) <0.0001
Sitagliptin 223 7.85 (0.05) -0.65 (0.05) -0.70 (-0.86, -0.54) <0.0001
1245.23 (met) (metformin background)
Placebo 207 7.90 (0.06) -0.13 (0.05) --- --- ---
Empagliflozin 10 mg 217 7.94 (0.05) -0.72 (0.05) -0.57 (-0.72, -0.42) <0.0001
Empagliflozin 25 mg 213 7.86 (0.06) -0.75 (0.06) -0.64 (-0.79, -0.48) <0.0001
1245.23 (met+SU) (metformin + sulphonylurea background)
Placebo 225 8.15 (0.06) -0.18 (0.05) --- --- ---
Empagliflozin 10 mg 225 8.07 (0.05) -0.80 (0.05) -0.64 (-0.79, -0.49) <0.0001
Empagliflozin 25 mg 216 8.10 (0.06) -0.77 (0.05) -0.60 (-0.76, -0.44) <0.0001
1245.19 (pioglitazone ± metformin background)
Placebo 165 8.16 (0.07) -0.14 (0.08) --- --- ---
Empagliflozin 10 mg 165 8.07 (0.07) -0.57 (0.07) -0.48 (-0.70, -0.26) <0.0001
Empagliflozin 25 mg 168 8.06 (0.06) -0.70 (0.07) -0.63 (-0.85, -0.41) <0.0001
The primary analysis was performed using an analysis of covariate model (ANCOVA) with baseline HbA1c as covariate, treatment and stratification factors for
randomization as fixed effects. Missing data was imputed using LOCF.

Five additional trials were deemed supportive of the indication by the applicant. For details
refer to Drs. Chong’s and Liu’s reviews. Briefly, improvement in glycemic control was
designated as a primary, secondary or tertiary objective in the following five trials and was
assessed after a variable duration of exposure shown in parentheses.

o Trial 1245.36 (24-weeks)- Randomized, double-blind, PLACEBO-controlled

trial to establish the efficacy and safety of 10 and 25 mg of empagliflozin
subjects with type 2 DM and renal impairment at baseline (all with eGFR <
90mL/min/1.73 m2).
o Trial 1245.48 (12-weeks)- Randomized, double-blind, PLACEBO-controlled trial
to establish the efficacy and safety of 10 and 25 mg of empagliflozin in adult
subjects with type 2 DM and hypertension.
o Trial 1245.25 (12, 52 and once yearly)-Randomized, double-blind, PLACEBO-
controlled cardiovascular outcomes trial in adult subjects with type 2 DM at
risk of CV-events. Assessment of glycemic control was a tertiary objective in a
trial not primarily designed to assess glycemic control.
o 1245.28 (52-weeks)-Randomized, double-blind, ACTIVE-controlled trial
comparing the efficacy and safety of 25 mg of empagliflozin to 1 to 4 mg of
glimepiride in adults with type 2 DM inadequately controlled on background
metformin.

Page 6 of 15

Reference ID: 3464120

 Division Director Review

o 1245.33 (18-weeks)-Randomized, double-blind, PLACEBO-controlled efficacy

trial to establish the efficacy and safety of 10 and 25 mg of empagliflozin used
as add-on to a fixed dose background basal insulin therapy in adult subjects
with type 2 DM.

Dr. Liu verified the primary analyses in trials 1245.28 and 1245.33. Empagliflozin 10 and 25
mg added to basal insulin (1245.33) resulted in a placebo adjusted mean change (97.5%) in
HbA1c from baseline of -0.62 (-0.82, -0.42) and-0.74 (-0.93, -0.56) respectively at 18 weeks
when used as an-add on to a fixed dose of basal insulin. Improvement in glycemic control
was similar between 25 mg of empagliflozin and glimepiride at the end of 52-weeks in trial
1245.33 [i.e., adjusted mean difference between treatment groups in the change in HbA1c
from baseline at Week 52 was -0.07% with a 97.5% CI of (-0.16, 0.02); pre-specified non-
inferiority margin 0.3%].

For each trial, changes in pre-specified key glycemic (fasting plasma glucose, responder
analyses) and non-glycemic secondary endpoints (change in systolic and diastolic blood
pressure, weight and insulin dose) were analyzed. Secondary endpoints were tested
according to pre-specified hierarchical testing sequences to control type-1 error across
multiple comparisons and two doses of empagliflozin. Results for analyses based on
secondary glycemic endpoints confirmed the findings based on HbA1c. Results for analyses
of non-glycemic endpoints confirm what is known about the class of SGLT-2 inhibitors. That
is, glucose lowering through this mechanism is associated with modest reductions in body
weight (~2.0 kg at 24-weeks) and systolic (~-3 to 4 mmHg) but not diastolic blood pressure. It
is unknown whether these changes meaningfully impact morbidity and mortality. Results for
these analyses are detailed in Dr. Liu’s and Chong’s reviews.

Dr. Liu integrated data across the four pivotal placebo-controlled trials and repeated the
primary efficacy analysis across various subgroups for the two empagliflozin doses.
Responses across the subgroups for race (refer to Figure 8 in Dr. Liu’s review) and geographic
region (refer to Figure 10 in Dr. Liu’s review) were similar to the overall response. Of note,
participants in the four pivotal trials were predominantly enrolled in Asia (~54%) and Asians
(56%) were the most represented racial group in these trials. Results of subgroup analyses by
geographical regions and race do not indicate differences in response across these
characteristic and suggest overall response is reasonably likely to reflect response for a US
population treated to US standard of care.

A significant interaction was noted between efficacy and baseline renal function for one of
the empagliflozin dosing group (i.e., 25 mg refer to Figure 14 in Dr. Liu’s review). Lesser
glucose lowering was observed for patients with higher degree of renal impairment at
baseline. This is expected from the drug’s mechanism of action which relies on renal function
and is similar to observations made in the Invokana and Farxiga programs.

Efficacy in Patients with Moderately Impaired Renal Function

Page 7 of 15

Reference ID: 3464120

 Division Director Review

The renal function subgroup analysis findings were confirmed in a dedicated trial (1245.36)
designed to compare the glycemic lowering effect of empagliflozin 25 mg against placebo at
the end of 24-weeks in subjects with type 2 DM and renal impairment at baseline. The LS
mean change in HbA1c from baseline was -0.6%, -0.5% and -0.2% for patients with an eGFR
between 60 to 90, 45 to 60 and 30 to 45 mL/min/1.73 m2 at baseline respectively. The
results shown below are modified from Table 25 in Dr. Chong’s review.

Table 2: Placebo Adjusted Change in HbA1c (%) from baseline to end-of-treatment (mITT
population using LOCF) in Subjects with eGFR < 90 mL/min/1.73 m2
LS mean
Treatment Baseline LS Mean
Study (Weeks) n difference (95% p-value
Arm Mean (SE) Change (SE)
CI)
eGFR ≥ 60 mL Empa 25 mg 97 7.96 (0.07) -0.63 (0.07) -0.68 (-0.49, -0.88) <0.0001
mL/min/1.73 m2 Placebo 95 8.09 (0.08) 0.06 (0.07)
eGFR 45 to < 60 Empa 25 mg 91 8.12 (0.09) -0.54 (0.07) -0.46 (-0.28, -0.56) <0.0001
2
mL/min/1.73 m Placebo 89 8.08 (0.09) -0.08 (0.08)
eGFR 30 to < 45 Empa 25 mg 96 7.95 (0.08) -0.21 (0.07) -0.39 (-0.19, -0.58) <0.0001
2
mL/min/1.73 m Placebo 98 8.01 (0.08) 0.17 (0.07)

8. Safety

Please refer to safety reviews by Drs. Chong, Charles and Senior for full details of the safety
findings in the empagliflozin application. Dr. Mahoney’s CDTL memorandum summarizes the
key safety findings in the application and the reader is referred to it for a synopsis. I agree
with the reviewers that the analyses of safety in the empagliflozin program do not raise
concerns that would preclude approval of the product. The major safety findings were
consistent with known and expected class-related adverse reactions. Adverse reactions
associated with empagliflozin use included adverse reactions related to volume contraction in
(e.g., hypotension and renal impairment), genital mycotic infections, increased urination, and
increased thirst. Some of these adverse reactions were more frequent and more pronounced
in specific patient subgroups. These adverse reactions are mitigated through product
labeling in the two approved products in the class.

Size of the Database Used in Safety Analyses:

For the review of overall safety Drs. Chong and Mahoney focused on a pool of 16 phase I, II
and III trials and their extensions (i.e., 1245.31 extensions of four pivotal studies and 1245.24
extensions of two phase-2 dose ranging studies ) referred to as SAF-5. This pool includes all
studies and trials carried out in individuals with type-2 diabetes mellitus. Studies in this pool
differed significantly in design, size, empagliflozin dose evaluated, duration (8 days to time to
event) and population enrolled (e.g., drug naïve, add-on to background therapy and enriched
for co-morbid conditions such as hypertension, renal impairment and high CV risk). SAF-5 is

Page 8 of 15

Reference ID: 3464120

 Division Director Review

the largest pool in terms of numbers and exposure duration. The 10 and 25 mg empagliflozin
doses account for most of the exposure in this pool. The trial contributing the most to the
pool in terms of numbers exposed and exposure duration was trial 1245.25 (CVOT trial). At
the time of submission all trials in SAF-5, were completed except for extensions of the four
pivotal trials (1245.31), trial 1254.28 (comparison to glimepiride) and trial 1245.25 (CVOT).
Database lock for the original submission was August 31st 2012. Dr. Chong also presents data
for the SAF-3 pool which is a subset of trials included SAF-5 and is limited to data from the
four placebo controlled pivotal efficacy trials and from their ongoing extensions.

Data in SAF-5 reflect exposure of 8400 patients to various doses of empagliflozin and a mean
empagliflozin exposure duration of 340 days. Demographic, anthropometric and disease
characteristics were balanced between empagliflozin and placebo treated patients. The
mean age of the entire SAF-5 population was 60 years and 6% were older than 75 years of
age. Sixty three percent (63%) of the subjects in SAF-5 were male; 62% were White, 34%
were Asian, and 4% were Black or African American. At baseline, 65% of the population had
had diabetes for 5 years or more. The mean HbA1c (SD) at baseline was 8.0% (0.84).
Baseline renal function based on estimated GFR was normal or mildly impaired in 83% of
patients and moderately impaired in 16% of patients (mean eGFR at baseline was 80
mL/min/1.73 m2).

Table 2: Exposure in SAF-5 pool (Source: Modified from Table 51 in Dr. Chong’s review).
SAF-5 Pool: ALL TYPE 2 DM TRIALS
(16 trials + 2 extension trials)

Subjects

Empagliflozin (all doses) 8400

10 mg 3630

25 mg 4602

Comparators (all) 4676

Placebo-only 3522
Patient-Years

Empagliflozin 7828
Comparators 4184
Placebo-only 2758

Page 9 of 15

Reference ID: 3464120

 Division Director Review

Deaths, Serious Adverse Events and Events Leading to Discontinuations:

Fewer incident deaths occurred on patients treated with empagliflozin (0.52 deaths per 100
patient years) than on placebo (1.04 deaths per 100 patient years) or comparator (0.78
deaths per 100 patient years). Dr. Chong reviewed narratives for all deaths in the program.
Drs. Chong and Mahoney conclude that clinical descriptions of the events in narratives do not
raise concerns for a drug-related specific cause of death.

The overall incidence of nonfatal serious adverse events was lower in empagliflozin-treated
patients (10.79 events per 100 patient-years) compared to placebo (16.54 events per 100
patient-years) and all comparators (12.78 events per 100 patient-years) (refer to Table 60 in
Dr. Chong’s review). More serious adverse events in the “Neoplasms benign, malignant and
unspecified (incl cysts and polyps)” and “Reproductive system and breast disorders” organ
classes were reported in empagliflozin-treated patients than in the placebo or comparator-
treated patients. Dr. Chong performed a causality assessment for events which were seen
more commonly in empagliflozin-treated patients. Drs. Chong and Mahoney both conclude
that narrative review do not raise concerns with regards to a potential causal relationship
between empagliflozin use and any specific serious adverse event identified in SAF-5.

The proportion of patients who discontinued due to occurrence of an adverse event was
similar between empagliflozin-treated (4.9%), placebo-treated (5.3%) and all comparator-
treated (4.8%) patients. More patients on empagliflozin (>2-fold) discontinued due to the
specific events of urinary tract infection and weight decreased (refer to Table 64 in Dr.
Chong’s review). A causal relationship to the drug is very likely in light of the marketed
observed imbalance, mechanism of action of the drug and observation of similar findings in
other members of the class.

Common Adverse Reactions

Adverse reactions attributed to empagliflozin use included events related to thirst, increased
urination and genital mycotic infections (refer to tables 124-127 in Dr. Chong’s review). For
the purpose of labeling none of the sponsor’s proposed pools are ideal. The applicant should
present pooled placebo-controlled data for the four pivotal trials up to the primary efficacy
endpoint as these were similar in design, duration, and dose and least subject to confounding
due to differential dropouts and use of rescue medication. In addition, preferred terms that
represent similar medical concepts (polyuria, polakiuria) should be combined for an accurate
representation of drug related adverse reactions.

Application Specific Concerns

Cardiovascular Risk Assessment:

Please refer to Drs. Chong and Charles reviews for details. To assess whether empagliflozin
use is associated with an unacceptable increased in cardiovascular (CV) risk, the applicant

Page 10 of 15

Reference ID: 3464120

 Division Director Review

performed a meta-analysis of clinical trials. This assessment of cardiovascular risk was

prospective, pre-specified, reviewed by the Agency and carried out in accordance with the
general principles laid out in the December 2008 FDA Guidance for Industry Diabetes
Mellitus-Evaluating Cardiovascular Risk in New Antidiabetic Therapies to Treat Type 2
Diabetes. The cardiovascular meta-analysis was based on data from six completed
randomized double-blind phase II and phase III trials (1245.19, 1245.20, 1245.23 met,
1245.23met+su, 1245.36, 1245.33) with treatment duration at least 12 weeks as well as data
from ongoing randomized double-blind phase III trials (1245.25, 1245.28 and 1245.31).

The primary comparison for the meta-analysis was between all empagliflozin doses and all
comparators. The primary safety endpoint was MACE+, a composite endpoint comprising CV
death, non-fatal myocardial infarction (MI), non-fatal stroke, or hospitalization for unstable
angina. A key secondary endpoint was MACE, a composite endpoint comprising CV death,
non-fatal MI, or non-fatal stoke. All events included in the meta-analysis were based on
positively adjudicated events determined by a blinded Clinical Event Committee using
accepted standardized event definitions. The pre-specified primary statistical analysis used a
Cox proportional hazards model, stratified by trial.

Results of the meta-analysis exclude the 1.8 pre-marketing CV-risk margin for MACE+ (b) (4)
events) and MACE-only (b) (4) events). Analyses of the individual components of MACE+ were
consistent with overall conclusions. For full details of these analyses the reader is referred to
Tables 12, 13 and 14 of Dr. Charles’s review.

Dr. Charles also performed subgroup analyses to explore CV-risk across baseline
characteristics including: gender, race, age, geographic region, BMI, smoking status, renal
function, duration of diabetes, and empagliflozin dose. Results of these analyses were
consistent with results based on overall data.

Liver Injury Risk Assessment:

Drs. Chong and Senior reviewed liver injury risk associated with empagliflozin use. To assess
liver injury risk, Dr. Chong considered reported adverse events in SAF-5, safety laboratory
monitoring in SAF-5 and performed a review of case narratives for participants with
laboratory evidence of hepatocellular injury severe enough to cause jaundice (e.g., alanine
amino transferase (ALT) ≥ 3 above the upper limit of normal (ULN) and total bilirubin (T.Bili) ≥
2-fold above the ULN or ALT ≥ 5 times the ULN).

Rates of reporting for adverse events preferred terms associated with abnormal liver
enzymes, jaundice, hepatitis, and hepatic failure were compared. No significant imbalance in
any preferred terms was observed with the exception of more frequent reporting of the term
‘hyperbilirubinemia’ (8/8400 vs. 0/4676) in patient using empagliflozin (refer to Table 94 in
Dr. Chong’s review).

Page 11 of 15

Reference ID: 3464120

 Division Director Review

Rates of significant hepatocellular injury defined as occurrence of a central laboratory

monitored ALT value ≥ 3, 5, 10 and 20 times the ULN detected at any time were compared in
SAF-5. A greater proportion of patients experienced an ALT ≥ 3 times the ULN in the
comparator group (0.51% versus 0.79% for empagliflozin versus all comparators). A greater
proportion of patient had an ALT ≥ 5 (0.19% versus 0.06%), 10 (0.08% versus 0) and 20 (0.01%
versus 0) times the upper limit of normal in the empagliflozin group (Refer to Table 96 in Dr.
Chong’s review).

All patients meeting central laboratory criteria suggestive of significant hepatic injury defined
as an ALT > 5 times the ULN or an ALT ≥ 3 times the ULN and a T.Bili > 2 times the ULN within
30 days of the noted ALT abnormality were reviewed by an independent adjudication
committee consisting of three hepatologists with expertise in drug induced liver injury (Drs.
Freston, Lewis and Watkins). These cases were also reviewed by Drs. Chong and Senior. For
the majority of cases an etiology other than empagliflozin was judged a more likely cause of
the liver event (refer to Table 95 in Dr. Chong’s review and Dr. Senior’s review). Many cases
identified as severe ALT elevation were not accompanied by a T.Bili rise, were not associated
with symptomatic complaints and therefore not investigated and resolved spontaneously in
spite of continued treatment. After review of these cases and the data above Drs. Chong,
Senior and Mahoney conclude that empagliflozin is unlikely to cause serious liver injury or
dysfunction but recommend continued evaluation in prospective ongoing studies.

Malignancy:

In the SAF-5 dataset, the proportion of participants who reported a malignant event (any
type) after at least six months of exposure to drug was balanced between groups (0.58%
versus 0.53%). Two specific malignant events (lung and melanoma) were reported more
frequently in patients exposed to empaglifozin for greater than six months.

In non-clinical studies empagliflozin was not mutagenic, clastogenic or carcinogenic at

clinically relevant exposures. In distribution studies, empagliflozin concentrations were not
elevated in lungs after oral dosing and clearance studies did not suggest drug accumulation in
any organs examined. Finally, non-clinical studies did not suggest a phototoxic risk associated
with oral empagliflozin use.

Seven patients exposed to empagliflozin for at least six months (0.08%) were reported to
have had a malignant lung neoplasm event compared to none in the comparator group. Time
between first exposure to empagliflozin and diagnosis ranged from 6 to 16 months. A
causality assessment performed by both Drs. Jennie Chang (Oncology consultant) and William
Chong revealed miscoding for one case (metastatic colon cancer), presence of confounders
(asbestos exposure, prior or current significant smoking exposure) for five cases, and
recurrence of a known cancer in one case. No dose response relationship was identified and
pathology reports do not indicate presence of a predominant cell type. The observed
incidence (150 cases per 100,000) observed in the empagliflozin arm is between the
estimated incidence for the general population (61 cases per 100,000) and high-risk

Page 12 of 15

Reference ID: 3464120

 Division Director Review

population (600 cases per 100,000). In totality, these data does not suggest a causal
relationship between empagliflozin use and lung cancer occurrence. Drs. Chong and
Mahoney do not believe this signal precludes approval and recommend continued follow-up
of lung neoplasm in long term clinical trials.

Six patients exposed to empagliflozin (0.07%) for at least six months were reported to have
had a melanoma event compared to none in the comparator group (0). Time between first
exposure to empagliflozin and diagnosis ranged from 7 to 12 months. Of these six patients,
two had a prior melanoma, one had multiple prior skin carcinomas, and one had “sun
damaged skin”. All patients were White. Non-melanoma skin cancer occurred more
frequently in comparators. No dose response was seen. Presence of prior disease and risk
factors in the majority of cases are not consistent with a drug related etiology. Drs. Chong
and Mahoney do not believe this signal precludes approval but recommend continued follow-
up of melanoma in long term clinical trials.

Other Significant Class-Related Adverse Reactions:

Adverse reactions related to volume contraction, impairment in renal function, increases in

hematocrit and increases in total and LDL cholesterol were observed with empagliflozin use.
Refer to Drs. Chong and Mahoney’s reviews for details. These are known class related
adverse reactions and magnitude of observed changes were consistent with what was
observed for the two other member of the class. Of note and in contrast to other programs,
acute change to serum creatinine (i.e., Week 1-3) after drug initiation was not captured in the
dedicated renal impairment study (first safety lab was evaluated at 12-weeks). These
reactions were for the most part not severe, would not preclude approval and can be
mitigated through product labelling.

9. Advisory Committee Meeting

No new efficacy or safety issue rose to the level of requiring the input from an advisory panel.
Therefore no advisory committee was convened.

10. Pediatrics
A pediatric study plan was submitted with the application. Pediatric studies required for
empagliflozin will be similar to pediatric studies required for other members of the class. The
applicant requests a waiver for children under 10 as type 2 DM rarely occurs in this age group
and a deferral for children 10 to 18 until a pediatric study plan is agreed to. The plan and
scientific rationale for waivers and deferrals were discussed with the Pediatric Review
Committee on February 12th 2014. The sponsor has proposed to carry out a PK/PD study
(1245.87) and an efficacy and safety study (1245.56). The Agency comments on the
acceptability of the plan have been issued.

Page 13 of 15

Reference ID: 3464120

 Division Director Review

11. Other Relevant Regulatory Issues

There are no other unresolved relevant regulatory issues

12. Labeling
The proprietary name of Jardiance has been found acceptable (Refer to OSE/DMEPA review
in DARRTS July 25th, 2013). This name will need to be subject to review when the application
is filed again.

Approval is not recommended and labeling negotiations will continue on the next cycle of
review.

13. Decision/Action/Risk Benefit Assessment

 Regulatory Action

Complete Response

 Risk Benefit Assessment

I recommend a Complete Response until the manufacturing site for both the drug substance
and product are determined to be compliant with current good manufacturing practices.

Otherwise, the applicant has demonstrated in adequate and well controlled trials that 10 and
25 mg of empagliflozin result in clinically meaningful improvement in long-term glycemic
control in patients with type 2 DM inadequately controlled with diet and exercise, metformin,
metformin and sulfonylurea and pioglitazone. Glucose lowering through SGLT-2 inhibition
was noted to be associated with a small placebo-adjusted reduction in weight and systolic
blood pressure. These effects were observed for the two approved products in the class and
are related to urinary glucose loss caused by SGLT-2 inhibition. Finally, the benefit of glucose
lowering using empagliflozin in patients with an eGFR ≤ 45 mL/min/1.73 m2 does not
outweigh the increased risk associated with volume contraction (e.g., hypotension and
worsening renal function).

Glucose lowering with empagliflozin is not associated with an inherently high risk of
hypoglycemia. The risk of hypoglycemia increases when empagliflozin is added to drugs
known to cause hypoglycemia (e.g., sulfonylurea and insulin). Drug-related risks identified in
the application (genital mycotic infection, hypovolemia related adverse reactions, increased
thirst and urination) are known risks associated with this class of glucose lowering agent.
Hypovolemic risks are expected to be dose-dependent as these results from urinary glucose
excretion (osmotic diuresis) and this pharmacodynamics effect was shown to be dose-

Page 14 of 15

Reference ID: 3464120

 Division Director Review

dependent at the two proposed doses in clinical pharmacology studies. Similar to

observations made in the canagliflozin and dapagliflozin programs, the elderly, patients with
renal impairment and patients on diuretics were particularly vulnerable to hypovolemia
related adverse reactions. The applicant’s pre-marketing CV-risk analysis excludes an excess
CV-risk of 1.8. Empagliflozin was found unlikely to cause significant liver dysfunction or
injury. The probability of a causal relationship between lung neoplasm, melanoma and
empagliflozin use was judged to be low after considering all available non-clinical and clinical
data.

 Recommendation for Postmarketing Risk Evaluation and Mitigation Strategies

These are deferred until the applicant resolves the above listed manufacturing deficiencies.

 Recommendation for other Postmarketing Requirements and Commitments

These are deferred until the applicant resolves the above listed manufacturing deficiencies.

Page 15 of 15

Reference ID: 3464120

 ---------------------------------------------------------------------------------------------------------
This is a representation of an electronic record that was signed
electronically and this page is the manifestation of the electronic
signature.
---------------------------------------------------------------------------------------------------------
/s/
----------------------------------------------------
JEAN-MARC P GUETTIER
03/04/2014

Reference ID: 3464120