The
new england journal
of
medicine
n engl j med 373;22 nejm.org November 26, 2015
2117
From the Lunenfeld-Tanenbaum Research
Institute, Mount Sinai Hospital (B.Z.)
and the Divisions of Endocrinology
(B.Z.) and Cardiology (D.F.), University
of Toronto — all in Toronto; the Depart-
ment of Medicine, Division of Nephrolo-
gy, Würzburg University Clinic, Würzburg
(C.W.), Boehringer Ingelheim Pharma,
Biberach (E.B., S.H.), and Boehringer
Ingelheim Pharma, Ingelheim (M.M.,
H.J.W., U.C.B.) — all in Germany; the Bio-
statistics Center, George Washington
University, Rockville, MD (J.M.L.); Boeh-
ringer Ingelheim Pharmaceuticals, Ridge-
field, CT (T.D.); Boehringer Ingelheim
Norway, Asker, Norway (O.E.J.); and the
Section of Endocrinology, Yale University
School of Medicine, New Haven, CT
(S.E.I.). Address reprint requests to Dr.
Zinman at Mount Sinai Hospital, 60 Mur-
ray St., Suite L5-024, Box 17, Toronto,
ONT M5T 3L9, Canada, or at zinman@
lunenfeld . ca.
This article was published on September 17,
2015, at NEJM.org.
N Engl J Med 2015;373:2117-28.
DOI: 10.1056/NEJMoa1504720
Copyright © 2015 Massachusetts Medical Society.
BACKGROUND
The effects of empagliflozin, an inhibitor of sodium–glucose cotransporter 2, in
addition to standard care, on cardiovascular morbidity and mortality in patients
with type 2 diabetes at high cardiovascular risk are not known.
METHODS
We randomly assigned patients to receive 10 mg or 25 mg of empagliflozin or
placebo once daily. The primary composite outcome was death from cardiovascu-
lar causes, nonfatal myocardial infarction, or nonfatal stroke, as analyzed in the
pooled empagliflozin group versus the placebo group. The key secondary compos-
ite outcome was the primary outcome plus hospitalization for unstable angina.
RESULTS
A total of 7020 patients were treated (median observation time, 3.1 years). The
primary outcome occurred in 490 of 4687 patients (10.5%) in the pooled empa-
gliflozin group and in 282 of 2333 patients (12.1%) in the placebo group (hazard
ratio in the empagliflozin group, 0.86; 95.02% confidence interval, 0.74 to 0.99;
P = 0.04 for superiority). There were no significant between-group differences in
the rates of myocardial infarction or stroke, but in the empagliflozin group there
were significantly lower rates of death from cardiovascular causes (3.7%, vs. 5.9%
in the placebo group; 38% relative risk reduction), hospitalization for heart failure
(2.7% and 4.1%, respectively; 35% relative risk reduction), and death from any
cause (5.7% and 8.3%, respectively; 32% relative risk reduction). There was no
significant between-group difference in the key secondary outcome (P = 0.08 for
superiority). Among patients receiving empagliflozin, there was an increased rate
of genital infection but no increase in other adverse events.
CONCLUSIONS
Patients with type 2 diabetes at high risk for cardiovascular events who received
empagliflozin, as compared with placebo, had a lower rate of the primary com-
posite cardiovascular outcome and of death from any cause when the study drug
was added to standard care. (Funded by Boehringer Ingelheim and Eli Lilly;
EMPA-REG OUTCOME ClinicalTrials.gov number, NCT01131676.)
ABSTRACT
Empagliflozin, Cardiovascular Outcomes,
and Mortality in Type 2 Diabetes
Bernard Zinman, M.D., Christoph Wanner, M.D., John M. Lachin, Sc.D.,
David Fitchett, M.D., Erich Bluhmki, Ph.D., Stefan Hantel, Ph.D.,
Michaela Mattheus, Dipl. Biomath., Theresa Devins, Dr.P.H.,
Odd Erik Johansen, M.D., Ph.D., Hans J. Woerle, M.D., Uli C. Broedl, M.D.,
and Silvio E. Inzucchi, M.D., for the EMPA-REG OUTCOME Investigators
Original Article
The New England Journal of Medicine
Downloaded from nejm.org on November 1, 2016. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
n engl j med 373;22 nejm.org November 26, 2015
2118
The
new england journal
of
medicine
T
ype 2 diabetes is a major risk factor
for cardiovascular disease,
1,2
and the pres-
ence of both type 2 diabetes and cardio-
vascular disease increases the risk of death.
3
Evidence that glucose lowering reduces the rates
of cardiovascular events and death has not been
convincingly shown,
4-6
although a modest cardio-
vascular benefit may be observed after a prolonged
follow-up period.
7
Furthermore, there is concern
that intensive glucose lowering or the use of
specific glucose-lowering drugs may be associat-
ed with adverse cardiovascular outcomes.
8
There-
fore, it is necessary to establish the cardiovascu-
lar safety benefits of glucose-lowering agents.
9
Inhibitors of sodium–glucose cotransporter 2
reduce rates of hyperglycemia in patients with
type 2 diabetes by decreasing renal glucose re-
absorption, thereby increasing urinary glucose
excretion.
10
Empagliflozin is a selective inhibitor
of sodium glucose cotransporter 2
11
that has
been approved for type 2 diabetes.
12
Given as
either monotherapy or as an add-on therapy, the
drug is reported to reduce glycated hemoglobin
levels in patients with type 2 diabetes, including
those with stage 2 or 3a chronic kidney dis-
ease.
13-20
Furthermore, empagliflozin is associ-
ated with weight loss and reductions in blood
pressure without increases in heart rate.
13-20
Empagliflozin also has favorable effects on mark-
ers of arterial stiffness and vascular resistance,
21
visceral adiposity,
22
albuminuria,
20
and plasma
urate.
13-19
Empagliflozin has been associated
with an increase in levels of both low-density
lipoprotein (LDL)
14
and high-density lipoprotein
(HDL) cholesterol.
13-16
The most common side
effects of empagliflozin are urinary tract infec-
tion and genital infection.
12
In the EMPA-REG OUTCOME trial, we exam-
ined the effects of empagliflozin, as compared
with placebo, on cardiovascular morbidity and
mortality in patients with type 2 diabetes at high
risk for cardiovascular events who were receiv-
ing standard care.
Methods
Study Oversight
The trial was designed and overseen by a steer-
ing committee that included academic investiga-
tors and employees of Boehringer Ingelheim. The
role of Eli Lilly was limited to cofunding the trial.
Safety data were reviewed by an independent aca-
demic data monitoring committee every 90 days
or at the discretion of the committee. Cardiovas-
cular outcome events and deaths were prospec-
tively adjudicated by two clinical-events commit-
tees (one for cardiac events and the other for
neurologic events), as recommended by the Food
and Drug Administration (FDA) guidelines.
9
A
list of investigators and committee members is
provided in Sections A and B, respectively, in the
Supplementary Appendix, which is available with
the full text of this article at NEJM.org.
The trial was conducted in accordance with
the principles of the Declaration of Helsinki and
the the International Conference on Harmonisa-
tion Good Clinical Practice guidelines and was
approved by local authorities. An independent
ethics committee or institutional review board
approved the clinical protocol at each participat-
ing center. All the patients provided written in-
formed consent before study entry.
All the authors were involved in the study
design and had access to the data, which were
analyzed by one of the study sponsors, Boeh-
ringer Ingelheim. All the authors vouch for the
accuracy and completeness of the data analyses
and for the fidelity of the study to the protocol,
available at NEJM.org. Members of the Univer-
sity of Freiburg conducted an independent statis-
tical analysis of cardiovascular outcomes (Section
B in the Supplementary Appendix). The manu-
script was drafted by the first and last authors
and revised by all the authors. Medical writing
assistance, which was paid for by Boehringer
Ingelheim, was provided by Fleishman-Hillard
Group.
Study Design
As described previously,
23
this was a randomized,
double-blind, placebo-controlled trial to assess
the effect of once-daily empagliflozin (at a dose
of either 10 mg or 25 mg) versus placebo on
cardiovascular events in adults with type 2 dia-
betes at high cardiovascular risk against a back-
ground of standard care. Patients were treated at
590 sites in 42 countries. The trial continued
until an adjudicated primary outcome event had
occurred in at least 691 patients.
Study Patients
Eligible patients with type 2 diabetes were adults
(≥18 years of age) with a body-mass index (the
weight in kilograms divided by the square of the
The New England Journal of Medicine
Downloaded from nejm.org on November 1, 2016. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
n engl j med 373;22 nejm.org November 26, 2015
2119
Empagliflozin in Type 2 Diabetes
height in meters) of 45 or less and an estimated
glomerular filtration rate (eGFR) of at least 30 ml
per minute per 1.73 m
2
of body-surface area, ac-
cording to the Modification of Diet in Renal
Disease criteria. All the patients had established
cardiovascular disease (as defined in Section C
in the Supplementary Appendix) and had received
no glucose-lowering agents for at least 12 weeks
before randomization and had a glycated hemo-
globin level of at least 7.0% and no more than
9.0% or had received stable glucose-lowering
therapy for at least 12 weeks before randomiza-
tion and had a glycated hemoglobin level of at
least 7.0% and no more than 10.0%. Other key
exclusion criteria are provided in Section D in
the Supplementary Appendix.
Study Procedures
Eligible patients underwent a 2-week, open-la-
bel, placebo run-in period in which background
glucose-lowering therapy was unchanged. Patients
meeting the inclusion criteria were then ran-
domly assigned in a 1:1:1 ratio to receive either
10 mg or 25 mg of empagliflozin or placebo once
daily. Randomization was performed with the use
of a computer-generated random-sequence and
interactive voice- and Web-response system and
was stratified according to the glycated hemo-
globin level at screening (<8.5% or ≥8.5%), body-
mass index at randomization (<30 or ≥30), renal
function at screening (eGFR, 30 to 59 ml, 60 to
89 ml, or ≥90 ml per minute per 1.73 m
2
), and
geographic region (North America [plus Austra-
lia and New Zealand], Latin America, Europe,
Africa, or Asia).
Background glucose-lowering therapy was to
remain unchanged for the first 12 weeks after
randomization, although intensification was per-
mitted if the patient had a confirmed fasting
glucose level of more than 240 mg per deciliter
(>13.3 mmol per liter). In cases of medical neces-
sity, dose reduction or discontinuation of back-
ground medication could occur. After week 12,
investigators were encouraged to adjust glucose-
lowering therapy at their discretion to achieve
glycemic control according to local guidelines.
Throughout the trial, investigators were encour-
aged to treat other cardiovascular risk factors
(including dyslipidemia and hypertension) to
achieve the best available standard of care accord-
ing to local guidelines. Patients were instructed
to attend the clinic at prespecified times, which
included a follow-up visit 30 days after the end
of treatment. Patients who prematurely discon-
tinued a study drug were to be followed for as-
certainment of cardiovascular outcomes, and
attempts were made to collect vital-status infor-
mation for any patient who was lost to follow-up,
as allowed by local guidelines.
Study Outcomes
The primary outcome was a composite of death
from cardiovascular causes, nonfatal myocardial
infarction (excluding silent myocardial infarc-
tion), or nonfatal stroke. The key secondary out-
come was a composite of the primary outcome
plus hospitalization for unstable angina. Defini-
tions of the major clinical outcomes are provid-
ed in Section E in the Supplementary Appendix.
Safety was assessed on the basis of adverse
events that occurred during treatment or within
7 days after the last dose of a study drug and
were coded with the use of the Medical Dictionary
for Regulatory Activities, version 18.0. Adverse events
of special interest included confirmed hypogly-
cemic adverse events (plasma glucose level, ≤70 mg
per deciliter [3.9 mmol per liter] or an event re-
quiring assistance), and adverse events reflecting
urinary tract infection, genital infection, volume
depletion, acute renal failure, bone fracture, dia-
betic ketoacidosis, and thromboembolic events.
Statistical Analysis
The primary hypothesis was noninferiority for
the primary outcome with empagliflozin (pooled
doses of 10 mg and 25 mg) versus placebo with
a margin of 1.3 for the hazard ratio.
9
We used a
four-step hierarchical-testing strategy for the
pooled empagliflozin group versus the placebo
group in the following order: noninferiority for
the primary outcome, noninferiority for the key
secondary outcome, superiority for the primary
outcome, and superiority for the key secondary
outcome.
Since interim data from the trial were included
in a new-drug application submitted to the FDA,
under the Haybittle–Peto rule, a two-sided P value
of 0.0498 or less was considered to indicate sta-
tistical significance in the final analyses.
23
For
the test of noninferiority for the primary out-
come with a margin of 1.3 at a one-sided level of
0.0249, at least 691 events were required to pro-
vide a power of at least 90% on the assumption
of a true hazard ratio of 1.0. Noninferiority for
The New England Journal of Medicine
Downloaded from nejm.org on November 1, 2016. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
n engl j med 373;22 nejm.org November 26, 2015
2120
The
new england journal
of
medicine
the primary outcome was determined if the up-
per boundary of the two-sided 95.02% confi-
dence interval was less than 1.3. Analyses were
based on a Cox proportional-hazards model, with
study group, age, sex, baseline body-mass index,
baseline glycated hemoglobin level, baseline eGFR,
and geographic region as factors. Estimates of
cumulative-incidence function were corrected
for death as a competing risk,
24
except for death
from any cause, for which Kaplan–Meier esti-
mates are presented. Because of the declining
numbers of patients at risk, cumulative-inci-
dence plots have been truncated at 48 months.
We calculated the number of patients who would
need to be treated to prevent one death on the
basis of the exponential distribution.
We performed the primary analysis using a
modified intention-to-treat approach among pa-
tients who had received at least one dose of a
study drug. Data for patients who did not have
an event were censored on the last day they were
known to be free of the outcome. Secondary
analyses included comparisons of the 10-mg dose
of empagliflozin versus placebo and the 25-mg
dose versus placebo. Sensitivity analyses are de-
scribed in the Section F in the Supplementary
Appendix. We analyzed the changes from base-
line in glycated hemoglobin level, weight, waist
circumference, systolic and diastolic blood pres-
sure, heart rate, LDL and HDL cholesterol, and
uric acid using a repeated-measures analysis as
a mixed model. Subgroup analyses are described
in Section F in the Supplementary Appendix.
Results
Study Patients
A total of 7028 patients underwent randomiza-
tion from September 2010 through April 2013.
Of these patients, 7020 were treated and in-
cluded in the primary analysis (Fig. S1 in Section
G in the Supplementary Appendix). Reasons for
premature discontinuation are provided in Table
S1 in Section H in the Supplementary Appendix.
Overall, 97.0% of patients completed the study,
with 25.4% of patients prematurely discontinu-
ing a study drug. Final vital status was available
for 99.2% of patients.
At baseline, demographic and clinical charac-
teristics were well balanced between the placebo
group and the empagliflozin group (Table S2 in
Section I in the Supplementary Appendix). Ac-
cording to the inclusion criteria, more than 99%
of patients had established cardiovascular dis-
ease, and patients were well treated with respect
to the use of lipid-lowering therapy and antihy-
pertensive medications at baseline. The median
duration of treatment was 2.6 years, and the
median observation time was 3.1 years; both
durations were similar in the pooled empa-
gliflozin group and the placebo group (Table S3
in Section J in the Supplementary Appendix).
Cardiovascular Outcomes
The primary outcome occurred in a significantly
lower percentage of patients in the empagliflozin
group (490 of 4687 [10.5%]) than in the placebo
group (282 of 2333 [12.1%]) (hazard ratio in the
empagliflozin group, 0.86; 95.02% confidence
interval [CI], 0.74 to 0.99; P<0.001 for noninferi-
ority and P = 0.04 for superiority) (Fig. 1A). The
key secondary outcome occurred in 599 of 4687
patients (12.8%) in the empagliflozin group and
333 of 2333 patients (14.3%) in the placebo
group (hazard ratio, 0.89; 95% CI, 0.78 to 1.01;
P<0.001 for noninferiority and P = 0.08 for supe-
riority).
As compared with placebo, empagliflozin
resulted in a significantly lower risk of death
from cardiovascular causes (hazard ratio, 0.62;
95% CI, 0.49 to 0.77; P<0.001) (Fig. 1B), death
from any cause (hazard ratio, 0.68; 95% CI, 0.57
to 0.82, P<0.001; Fig. 1C), and hospitalization
for heart failure (hazard ratio, 0.65; 95% CI, 0.50
to 0.85; P = 0.002) (Fig. 1D). Hazard ratios for
cardiovascular outcomes with empagliflozin ver-
sus placebo are shown in Table 1. Absolute re-
ductions in incidence rates for cardiovascular
outcomes are provided in Table S4 in Section K
in the Supplementary Appendix. All categories
of death from cardiovascular causes contributed
to the reduction in cardiovascular death in the
empagliflozin group (Table S5 in Section L in
the Supplementary Appendix). There were no
significant between-group differences in the oc-
currence of myocardial infarction or stroke (Ta-
ble 1). Myocardial infarction was reported in
4.8% of patients in the empagliflozin group and
5.4% of those in the placebo group, and stroke
in 3.5% and 3.0% of patients, respectively.
For the primary and key secondary outcomes,
hazard ratios for the comparison between the
10-mg dose of empagliflozin versus placebo and
the 25-mg dose versus placebo were virtually
The New England Journal of Medicine
Downloaded from nejm.org on November 1, 2016. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
n engl j med 373;22 nejm.org November 26, 2015
2121
Empagliflozin in Type 2 Diabetes
Figure 1. Cardiovascular Outcomes and Death from Any Cause.
Shown are the cumulative incidence of the primary outcome (death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke) (Panel A), cumulative incidence of
death from cardiovascular causes (Panel B), the Kaplan–Meier estimate for death from any cause (Panel C), and the cumulative incidence of hospitalization for heart failure (Panel D)
in the pooled empagliflozin group and the placebo group among patients who received at least one dose of a study drug. Hazard ratios are based on Cox regression analyses.
Patients with Event (%)
20
15
5
10
0
0
126 18 24
30
36
42
48
Month
A Primary Outcome
No. at Risk
Empagliflozin
Placebo
4687
2333
4580
2256
4455
2194
4328
2112
3851
1875
2821
1380
2359
1161
1534
741
370
166
D Hospitalization for Heart Failure
C Death from Any Cause
B Death from Cardiovascular Causes
Hazard ratio, 0.86 (95.02% CI, 0.74–0.99)
P=0.04 for superiority
Placebo
Empagliflozin
Patients with Event (%)
9
7
8
3
4
5
6
0
1
2
0
126 18 24
30
36
42
48
Month
No. at Risk
Empagliflozin
Placebo
4687
2333
4651
2303
4608
2280
4556
2243
4128
2012
3079
1503
2617
1281
1722
825
414
177
Hazard ratio, 0.62 (95% CI, 0.49–0.77)
P<0.001
Placebo
Empagliflozin
Patients with Event (%)
15
5
10
0
0
126 18 24
30
36
42
48
Month
No. at Risk
Empagliflozin
Placebo
4687
2333
4651
2303
4608
2280
4556
2243
4128
2012
3079
1503
2617
1281
1722
825
414
177
Hazard ratio, 0.68 (95% CI, 0.57–0.82)
P<0.001
Placebo
Empagliflozin
Patients with Event (%)
7
3
4
5
6
0
1
2
0
126 18 24
30
36
42
48
Month
No. at Risk
Empagliflozin
Placebo
4687
2333
4614
2271
4523
2226
4427
2173
3988
1932
2950
1424
2487
1202
1634
775
395
168
Hazard ratio, 0.65 (95% CI, 0.50–0.85)
P=0.002
Placebo
Empagliflozin
The New England Journal of Medicine
Downloaded from nejm.org on November 1, 2016. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
