EMPA-REG OUTCOME Results in Detail: Numbers, Subgroups, and Time Course

At a glance

• Trial: EMPA-REG OUTCOME (BI 10773 [Empagliflozin] Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients)
• N: 7,020 randomized (2,333 empagliflozin 10 mg; 2,345 empagliflozin 25 mg; 2,333 placebo)
• Intervention: Empagliflozin 10 mg or 25 mg once daily
• Comparator: Placebo, added to standard of care
• Duration: Median observation 3.1 years
• Primary endpoint: Time to first occurrence of 3-point MACE (cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke)
• Key result: HR 0.86 (95.02% CI, 0.74-0.99; p = 0.04 for superiority)

The Primary Composite: What the 14% Actually Means

The primary analysis pooled both empagliflozin doses against placebo. In the pooled empagliflozin group, 490 of 4,687 patients (10.5%) experienced a MACE event compared with 282 of 2,333 (12.1%) in the placebo group. The absolute risk reduction was 1.6 percentage points over the median 3.1-year follow-up, translating to a number needed to treat (NNT) of approximately 63 per year.

The trial used a hierarchical testing procedure. Empagliflozin first had to demonstrate noninferiority (upper bound of the 97.5% CI for the hazard ratio <1.3), which it cleared easily. Only then was the superiority test performed using the remaining alpha, which is why the confidence interval is reported as 95.02% rather than the conventional 95%.

One detail often overlooked: the event rate in the placebo arm (12.1% over 3.1 years) was lower than the 15% rate projected during trial design. Despite this, the trial reached statistical significance for superiority, a fact that speaks to the consistency of the treatment effect.

Individual MACE Components: Where the Signal Lives

The three-point MACE composite hides a dramatic imbalance among its components. The table below breaks down each element from the primary publication.

HealthRX.com MACE Component Dissection Framework

Two patterns demand attention. First, cardiovascular death drove the entire composite. The 38% relative risk reduction (HR 0.62) was far larger than the overall 14% MACE reduction, meaning the composite actually diluted the mortality signal. Second, nonfatal stroke trended numerically higher with empagliflozin (HR 1.24), though the confidence interval crossed 1.0. This stroke signal did not replicate in subsequent SGLT2 inhibitor trials (CANVAS, DECLARE-TIMI 58), and most experts consider it a chance finding in a secondary endpoint.

The heart failure hospitalization result (HR 0.65, 95% CI 0.50-0.85) was a prespecified secondary endpoint. This 35% relative risk reduction arrived before dedicated heart failure trials of SGLT2 inhibitors existed and was a major catalyst for the DAPA-HF and EMPEROR-Reduced programs that followed.

Time-Course Pattern: Separation Within Months

The Kaplan-Meier curves for cardiovascular death and heart failure hospitalization began separating from placebo within the first 2 to 3 months. This rapid onset was considered inconsistent with an atherosclerosis-mediated mechanism (plaque stabilization or regression typically requires 12 to 18 months to produce detectable event reduction). Instead, the early separation pointed toward hemodynamic mechanisms: reduced preload, decreased arterial stiffness, or direct effects on myocardial energetics.

For the overall MACE composite, separation was more gradual and became clearly apparent by approximately 6 months, with the gap widening steadily through the trial's duration. The mortality curves continued to diverge without any sign of attenuation through the end of the observation period, suggesting a sustained rather than transient benefit.

This time-course pattern had direct implications for clinical practice. It suggested that patients at highest short-term risk for heart failure decompensation or cardiovascular death could benefit quickly after initiation, rather than requiring months of therapy before any protective effect emerged.

Subgroup Analyses: Consistency and Exceptions

The investigators performed subgroup analyses across 16 prespecified baseline variables. The cardiovascular death reduction was broadly consistent regardless of age (<65 vs. ≥65), sex, race, BMI category, baseline HbA1c (<8.5% vs. ≥8.5%), eGFR strata (≥60 vs. 30-59 mL/min/1.73m²), and use of insulin, metformin, or sulfonylureas at baseline.

A few notable subgroup observations:

• eGFR 30-59 mL/min/1.73m²: The cardiovascular benefit was maintained even in patients with moderate renal impairment, despite the reduced glucose-lowering efficacy of empagliflozin at lower GFR. This reinforced the concept that the cardiovascular mechanism was at least partially glucose-independent.
• Heart failure history at baseline: Approximately 10% of participants had a history of heart failure. The heart failure hospitalization benefit was consistent in patients with and without baseline heart failure, though the absolute risk reduction was larger in those with prior heart failure.
• Race and region: No significant interaction by geographic region (North America, Europe, Asia, Latin America, Africa) or racial group was observed for the primary endpoint.
• Dose comparison: The 10 mg and 25 mg empagliflozin groups showed similar hazard ratios for the primary composite (HR 0.85 and 0.86, respectively), with overlapping confidence intervals. The trial was not powered for dose-response comparison, and the FDA-approved labeling reflects this by granting the cardiovascular indication to both doses.

Glycemic and Metabolic Secondary Endpoints

Beyond cardiovascular events, the trial captured metabolic parameters that contextualize the overall clinical picture.

The HbA1c reduction was modest (0.38% adjusted difference) and attenuated over time as investigators were permitted to intensify background glucose-lowering therapy. The blood pressure reduction of approximately 3/1 mmHg was immediate and sustained. Weight loss of about 1.2 kg over placebo was also observed, though this is less than the 2 to 3 kg typically seen in shorter glucose-lowering trials with SGLT2 inhibitors, possibly because of compensatory caloric intake over the longer follow-up.

Mediation analyses published subsequently estimated that blood pressure and HbA1c changes together could explain only a small fraction of the observed cardiovascular mortality benefit, reinforcing that the drug's protective mechanism operates through pathways beyond classical risk-factor modification.

Renal Outcomes: A Signal That Launched a New Indication

While the primary trial focused on cardiovascular endpoints, a prespecified composite renal outcome (incident or worsening nephropathy) showed a 39% relative risk reduction with empagliflozin (HR 0.61, 95% CI 0.53-0.70). This composite included progression to macroalbuminuria, doubling of serum creatinine, initiation of renal replacement therapy, or renal death.

The renal data from EMPA-REG OUTCOME, combined with similar signals from CREDENCE (canagliflozin) and later EMPA-KIDNEY, established SGLT2 inhibitors as a standard renal-protective therapy. Current KDIGO 2024 guidelines recommend SGLT2 inhibitors for patients with CKD and an eGFR ≥20 mL/min/1.73m², with or without diabetes.

Safety Results Worth Noting

Empagliflozin showed increased rates of genital mycotic infections (6.4% vs. 1.8% in women; 1.8% vs. 0.4% in men). Urinary tract infections were not significantly different between groups.

The rate of diabetic ketoacidosis was very low in both arms (less than 0.1%) and not significantly different, though the trial predated widespread clinical awareness of euglycemic DKA with SGLT2 inhibitors. No increased fracture risk was observed, in contrast to a signal later seen with canagliflozin in CANVAS.

Volume depletion events were modestly more frequent with empagliflozin (5.1% vs. 4.2%), consistent with the osmotic diuretic mechanism.

Limitations the Authors Acknowledged

The EMPA-REG OUTCOME investigators and the broader community identified several constraints on interpretation:

1. Enrolled population: All patients had established atherosclerotic cardiovascular disease. The results could not be extrapolated to primary prevention T2D populations. The subsequent DECLARE-TIMI 58 trial (dapagliflozin) enrolled a broader population and showed a smaller, non-significant MACE reduction.
2. Stroke signal: The numerical increase in nonfatal stroke (HR 1.24) remained unexplained. While likely a chance finding, it precluded definitive conclusions about cerebrovascular safety.
3. Pooled dose analysis: The primary analysis pooled both empagliflozin doses. True dose-response assessment was not feasible with the sample size.
4. Background therapy changes: Open-label glucose-lowering intensification was permitted, which may have attenuated glycemic separation and potentially blunted event-rate differences over time.
5. Median follow-up: At 3.1 years, the trial cannot speak to very long-term outcomes beyond that window.

How EMPA-REG Changed Practice

The trial's results directly shaped the 2019 ESC/EASD guidelines and the ADA Standards of Care, which moved SGLT2 inhibitors (alongside GLP-1 receptor agonists) to first-line add-on therapy after metformin for patients with T2D and established cardiovascular disease. More recent ADA guidance positions these agents as first-line regardless of metformin use.

The FDA granted empagliflozin a cardiovascular death risk reduction indication in December 2016, making it the first diabetes drug to carry such a label claim based on a dedicated outcomes trial.

Frequently asked questions

›What was the primary endpoint result in EMPA-REG OUTCOME?

The pooled empagliflozin group showed a 14% relative risk reduction in 3-point MACE (cardiovascular death, nonfatal MI, nonfatal stroke) compared with placebo (HR 0.86, 95.02% CI 0.74-0.99, p = 0.04 for superiority).

›How large was the cardiovascular death reduction with empagliflozin?

Empagliflozin reduced cardiovascular death by 38% relative to placebo (HR 0.62, 95% CI 0.49-0.77, p <0.001). This was the single largest driver of the primary composite endpoint result.

›How quickly did the treatment effect appear in EMPA-REG OUTCOME?

Kaplan-Meier curves for cardiovascular death and heart failure hospitalization separated from placebo within approximately 2 to 3 months, suggesting a hemodynamic rather than atherosclerosis-modifying mechanism.

›Did EMPA-REG OUTCOME show a benefit for heart failure?

Yes. Heart failure hospitalization was reduced by 35% (HR 0.65, 95% CI 0.50-0.85, p = 0.002). This finding was instrumental in launching subsequent dedicated heart failure trials of SGLT2 inhibitors.

›Was the stroke result in EMPA-REG a safety concern?

Nonfatal stroke showed a non-significant numerical increase with empagliflozin (HR 1.24, 95% CI 0.92-1.67). This did not replicate in other SGLT2 inhibitor cardiovascular outcomes trials and is generally regarded as a chance finding.

›Did the 10 mg and 25 mg empagliflozin doses differ in cardiovascular benefit?

No meaningful difference was detected. The 10 mg group had an HR of 0.85 and the 25 mg group had an HR of 0.86 for the primary composite, with overlapping confidence intervals. The trial was not powered for formal dose-response testing.

›What were the main side effects in EMPA-REG OUTCOME?

Genital mycotic infections were significantly more common with empagliflozin (6.4% vs. 1.8% in women). Volume depletion events were modestly increased (5.1% vs. 4.2%). Diabetic ketoacidosis, fractures, and urinary tract infections were not significantly different from placebo.

›Did patients with kidney disease benefit from empagliflozin in this trial?

The cardiovascular benefit was maintained in patients with eGFR 30-59 mL/min/1.73m². A prespecified renal composite endpoint showed a 39% relative risk reduction (HR 0.61), which helped spark the development of dedicated renal outcomes trials for the SGLT2 inhibitor class.

›How did EMPA-REG OUTCOME change diabetes treatment guidelines?

The trial led to SGLT2 inhibitors being recommended as preferred add-on therapy for T2D patients with established cardiovascular disease in ESC/EASD, ADA, and other major guidelines. It also resulted in the first cardiovascular death reduction labeling for any diabetes drug.

›Can EMPA-REG results be applied to patients without cardiovascular disease?

The trial enrolled only patients with established atherosclerotic cardiovascular disease. DECLARE-TIMI 58 later tested dapagliflozin in a broader population and did not show a significant MACE reduction, though heart failure hospitalization benefits were preserved. Extrapolation to primary prevention requires caution.

References

1. Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015;373(22):2117-2128. PubMed
2. FDA. Jardiance (empagliflozin) prescribing information. AccessData
3. McMurray JJV, Solomon SD, Inzucchi SE, et al. Dapagliflozin in patients with heart failure and reduced ejection fraction. N Engl J Med. 2019;381(21):1995-2008. PubMed
4. Perkovic V, Jardine MJ, Neal B, et al. Canagliflozin and renal outcomes in type 2 diabetes and nephropathy. N Engl J Med. 2019;380(24):2295-2306. PubMed
5. Cosentino F, Grant PJ, Aboyans V, et al. 2019 ESC Guidelines on diabetes, pre-diabetes, and cardiovascular diseases. Eur Heart J. 2020;41(2):255-323. 2019 ESC Guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD
6. American Diabetes Association. Standards of Care in Diabetes, 2023. Diabetes Care. 2023;46(Suppl 1). PubMed