The Heart of the Matter – Is EMPA-REG a Game Changer for Diabetes Management?


Written By

Amanda Schartel-Lasota, Pharm.D., BCACP

Reviewed By

Daniel Longyhore, Pharm.D., BCACP
Stuart T. Haines, Pharm.D., BCPS, BCACP


Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015 Sept 17 [Epub ahead of print].

When deciding which medications to use to treat a patient with type 2 diabetes, there are many options after initial treatment with metformin. The selection of a second-line agent is based on many factors including efficacy, cost, adverse effects, effect on weight, comorbidities, hypoglycemia risk, and patient preference.1  No clear consensus exists.  One characteristic that would truly help differentiate agents is their effect on cardiovascular events.


Many studies have shown the benefits of controlling blood pressure and using lipid-lowering therapy (particularly statins) to reduce cardiovascular disease (CVD) in patients with diabetes.2 However, CVD remains the number one cause of death in people with diabetes, with rates about twice as high as the general population.2,3 The United Kingdom Prospective Diabetes Study (UKPDS) was the first landmark trial to demonstrate a reduction in diabetes related-complications including cardiovascular events in patients who took metformin.4 Similarly, the PROactive study found a reduction in cardiovascular events in patients using pioglitazone.5  Both of these studies have significant limitations.  It could be argued that no antidiabetic medication has demonstrated a clear cardiovascular benefit and some anti-diabetes drugs have even been linked to worse cardiovascular outcomes (rosiglitazone, saxagliptin).1 The newest class of antidiabetic medications called sodium-glucose cotransporter (SGLT) 2 inhibitors work to lower blood glucose by inhibiting reabsorption of glucose in the kidneys and increasing urinary excretion.  These agents include canagliflozin, dapagliflozin, and empagliflozin.7 Could these agents reduce the rate of cardiovascular events?


The Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes (EMPA-REG OUTCOME) study sought to determine the effects of empagliflozin on cardiovascular morbidity and mortality in patients with type 2 diabetes at high risk for CVD.  The study was a randomized, double blind, placebo-controlled, multicenter trial that enrolled adult patients with type 2 diabetes and at high risk for cardiovascular events (Table 1) with a BMI ≤ 45 kg/m2 and eGFR ≥ 30 ml/min/1.73 m2.  For those who had not previously received anti-diabetes medications, the patient’s A1c was between 7 and 9% at the time of enrollment.  For those on treatment at the time of enrollment, the patient’s A1c could be between 7 and 10% if the medication regimen was stable.  Major exclusion criteria included acute coronary syndrome, stroke, or transient ischemic attack within 2 months prior to enrollment; planned cardiac surgery or angioplasty within 3 months; and fasting blood glucose > 240 mg/dL.  The primary outcome was the composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke.  The secondary outcome was a composite of the primary outcome and hospitalization for unstable angina.  This study used a four-step hierarchical-testing strategy to first determine noninferiority of the outcomes and then to determine superiority.7


Table 1. Definition of High Risk for Cardiovascular Events

Presence of 1 or more of the following:

  • History of myocardial infarction >2 months prior to informed consent
  • Evidence of multi-vessel coronary artery disease i.e. in ≥2 major coronary arteries or the left main coronary artery, documented by any of the following:
    • Presence of significant stenosis: ≥50% luminal narrowing during angiography (coronary or multi-slice computed tomography)
    • Previous revascularization (percutaneous transluminal coronary angioplasty ± stent or coronary artery bypass graft >2 months prior to consent
    • The combination of revascularization in one major coronary artery and significant stenosis (≥50% luminal narrowing) in another major coronary artery
  • Evidence of single-vessel coronary artery disease, ≥50% luminal narrowing during angiography (coronary or multi-slice computed tomography) not subsequently successfully revascularized, with at least 1 of the following:
    • A positive non-invasive stress test for ischemia
    • Hospital discharge for unstable angina ≤12 months prior to consent
  • Unstable angina >2 months prior to consent with evidence of single- or multi-vessel coronary artery disease
  • History of stroke (ischemic or hemorrhagic) >2 months prior to consent
  • Occlusive peripheral artery disease documented by any of the following:
    • Limb angioplasty, stenting, or bypass surgery
    • Limb or foot amputation due to circulatory insufficiency
    • Evidence of significant peripheral artery stenosis (>50% on angiography, or >50% or hemodynamically significant via non-invasive methods) in 1 limb
    • Ankle brachial index <0.9 in ≥1 ankle


The study enrolled 7020 patients who were randomized and stratified by baseline A1c, BMI, renal function, and geographic location.  Patients received empagliflozin 10 mg (n = 2345), empagliflozin 25 mg (n = 2342), or placebo (n = 2333).  Additionally, patients were treated for other cardiovascular risk factors based on local standards of care and clinical practice guidelines.  Baseline characteristics between the groups were similar.  The “average” patient in the study was 63 years old, white (71%), and male (72%) with a baseline A1c of 8%.  The majority of patients had pre-existing coronary artery disease (76%) and were being treated with metformin (74%), an ACEI or ARB (81%), a beta-blocker (65%), a statin (77%), and aspirin (83%).8


After a median of 2.6 years of treatment and 3.1 years of observed follow-up, empaglifozin significantly reduced the rate of the composite primary outcome when compared to placebo (10.5% vs. 12.1%; HR 0.86 [95% CI 0.74-0.99]; p<0.001 for noninferiority and p=0.04 for superiority) with a NNT of 63.7  The secondary composite outcome was found to be non-inferior (12.8% vs. 14.3%; HR 0.89 [95% CI 0.78-1.01]; p<0.001, but not superior; p=0.08).  The study also found significant reductions in rates of several individual end-points including all-cause mortality, death from cardiovascular causes, and hospitalizations for heart failure.  The rates of myocardial infarction or stroke were not significantly different.  Of note, the rate of fatal and non-fatal stroke was numerically greater in the empagliflozin group versus placebo (3.5% vs. 3%; p=0.26).  Overall, the rates of adverse events, serious adverse events, and adverse events leading to discontinuation of study drug were similar among the groups.7 There were statistically higher rates of genital infections in both men (5% vs. 1.5%) and women (10% vs. 2.6%) with empagliflozin (p<0.001),7 which is consistent with the known adverse event profile of SGLT 2 inhibitors.7 There were also significantly higher rates of acute renal failure and acute kidney injury in the placebo group.7


This study demonstrates that empagliflozin has a positive effect on cardiovascular outcomes in patients with type 2 diabetes at high risk for CVD.  The benefits were largely due to reductions in deaths from cardiovascular causes and hospitalizations for heart failure.  The subgroup analyses showed that patients ≥65 years old and those with baseline A1c values <8.5% received the greatest benefits.  The results of this study are strengthened by its trial design (randomized, double-blinded, placebo-controlled), large sample size, and the ability of investigators to control for other CV risk factors by using clinical practice guideline-driven therapy.  Although the internal validity of the study is very strong, the results are somewhat limited by the disproportionate number of patients who were white, which is in contrast to the higher prevalence of type 2 diabetes among blacks in the United States.9 Furthermore, the numerical increase in rate of stroke seen in the empagliflozin group is somewhat concerning and additional studies will be needed to explore this potential risk.


The results of EMPA-REG OUTCOME are particularly encouraging when you compare them to previous studies that show CV benefit of medications used to treat diabetes.  In UKPDS, metformin was shown to reduce all-cause mortality and myocardial infarctions.  However, this was seen in a subpopulation of overweight patients and only a few hundred subjects were enrolled in this part of the study (342 in the metformin group, 411 in the diet group).  Furthermore, when metformin was added to a sulfonylurea, patients had a significant increase in death (all-cause and diabetes-related).4 The PROactive study did find a significant reduction in the secondary composite endpoint (all-cause mortality, non-fatal myocardial infarction, stroke) with pioglitazone, but this benefit was largely offset by the significant increase in new-onset heart failure and heart-failure hospitalization.5


The compelling evidence from the EMPA-REG OUTCOME study coupled with the other benefits of SGLT2 inhibitors (reduction in weight and blood pressure, mild adverse effects) should make empagliflozin the preferred second-line agent from the treatment of type 2 diabetes.  Perhaps the one caveat to this recommendation would be for those patients at high risk for UTIs or genital infections.  While EMPA-REG OUTCOME is the only completed study evaluating the effects of a SGLT2 inhibitor on CVD, the CANVAS study is a randomized, double blind, placebo-controlled trial currently underway.  Similar to the EMPA-REG OUTCOME study, CANVAS is evaluating the effects of canagliflozin on cardiovascular events (CV death, myocardial infarction, and stroke) in patients with uncontrolled type 2 diabetes and either a history of CV events or at high risk for CV events.9 The results of the CANVAS study will help clarify whether the cardiovascular benefits are specific to empagliflozin or a class effect.  What do you think? Should we use empagliflozin as the preferred add-on therapy for patients whose diabetes is not adequately controlled by metformin alone?



Dan Gillis's picture

"The compelling evidence from the EMPA-REG OUTCOME study ...should make empagliflozin the preferred second-line agent from the treatment of type 2 diabetes." One needs to be careful not to generalize beyond the population studied...patients with established CVD (also 70% white and male). Application beyond a similar population to the point of changing standards of care is premature. Also of note, the findings were not significant in the sub-population with established HF. I find it troubling that in presenting the results, e.g., AHA a couple of weeks ago, the lead author has stated that this was due to sample size and asserts that there is a benefit in this population as well. That could be, but just as well may not be; it remains to be proven. To state it as fact is scientifically irresponsible and disappointing coming from a thought leader such as Dr. Inzucchi. EMPA-REG puts SGLT2I in a favorable light for many patients, but it's not a game changer. Individualization remains key
Lucas G Hill's picture

I appreciate the author's willingness to embrace compelling new evidence given a tendency for clinicians to hang onto the past (e.g. anticoagulation). As the value of various add-ons have been called into question, and the durability of sulfonylureas seems particularly limited, it is tempting to act on this study. My primary reservations are related to cost and genitourinary infections. Practicing in an FQHC, a limited formulary of 340B, bulk PAP, and cheap medications are at my disposal. I could utilize dapagliflozin but can't say with certainty that reduced CV risk is a class effect. Regarding GU infections, how am I to determine whether a patient is at "high risk" for these? An incidence of 10% in tightly-controlled studies is concerning. Finally, the industry-sponsored nature of this study leaves me with only guarded optimism. While we typically rely on industry to fund prospective studies of the size necessary to assess CV outcomes, I can't help but worry about the influence of $.

I'm curious if anyone has thoughts on the Categories of Cardiovascular Death (Table S5 in the supplementary appendix). If you note in the causes of death (also as indicated by the study authors) there is no difference in death fro stroke or acute MI (and no difference in cardiogenic shock). There are differences in CV sudden death (0.3% difference between groups), worsening heart failure (0.5%), and "other cardiovascular death" (a whopping 0.8%). They indicate "other cardiovascular death" includes fatal cases of death that were not assessable due to lack of information and were PRESUMED (emphasis mine) to be CV deaths... This table puzzles me for two reasons: 1) why would investigators assume all deaths without identifiable cause to be for cardiovascular reasons? 2) why would investigators bury a small table which effectively tabulates your primary outcome in your appendix? What do others think? Am I missing something obvious? Is there a possible threat to validity here?