Principal Findings:

The primary outcome, cardiovascular (CV) death or HF hospitalization, for empagliflozin vs. placebo, was 13.8% vs. 17.1% (hazard ratio [HR] 0.79, 95% confidence interval [CI] 0.69-0.90, p < 0.001).

• CV death: 7.3% vs. 8.2% (HR 0.91, 95% CI 0.76-1.09)
• HF hospitalization: 8.6% vs. 11.8% (HR 0.71, 95% CI 0.60-0.83)
• For the primary outcome, the benefit was similar among patients with or without type 2 diabetes. The benefit appeared somewhat attenuated among patients with EF ≥60%.

Secondary outcomes for empagliflozin vs. placebo:

• Total hospitalizations: 407 vs. 541 (p < 0.001)
• Change in mean eGFR slope/year: -1.25 vs. -2.62 (p < 0.001)
• Composite renal outcome 3.6% vs. 3.7% (p > 0.05)
• All-cause mortality: 13.4% vs. 14.2% (HR 0.92, 95% CI 0.77-1.10, p > 0.05)
• New-onset type 2 diabetes among patients with prediabetes: 12.0% vs. 14.0% (p > 0.05)

Patients with and without type 2 diabetes: Type 2 diabetes was present in 49% of patients. Empagliflozin reduced the rate of the primary outcome irrespective of diabetes status (with diabetes: HR 0.79 [95% CI 0.67-0.94]; without diabetes: HR 0.78 [95% CI 0.64-0.95]; p for interaction = 0.92). The effect of empagliflozin to attenuate eGFR decline during double-blind treatment was also present in patients with and without diabetes, although more pronounced in patients with diabetes (1.77 in diabetes vs. 0.98 mL/min/1.73 m² in patients without diabetes; p for interaction = 0.01).

Pooled analysis of EMPEROR-Reduced and EMPEROR-Preserved on renal outcomes (profound and sustained decreases in eGFR or renal replacement therapy), total n = 9,718: 2.8% vs. 3.5% for empagliflozin vs. placebo, with significant heterogeneity between both trials (p = 0.016 for interaction).

Outcomes based on EF (33% had EF 41-49%; remainder had EF ≥50%):

• Primary endpoint for empagliflozin vs. placebo for EF ≥50%: 6.7% vs. 8.0% (p = 0.024); for EF 41-49%: 7.2% vs. 10% (p = 0.002; p for interaction = 0.27)
• Total HF hospitalizations for EF ≥50%: 4.5% vs. 5.7% (p = 0.013); for EF 41-49%: 3.8% vs. 6.5% (p < 0.001; p for interaction = 0.06)
• All-cause mortality for EF ≥50%: 6.1% vs. 6.1% (p = 0.84); EF 41-49%: 7.7% vs. 8.0% (p = 0.72; p for interaction > 0.05)
• Change from baseline in Kansas City Cardiomyopathy Questionnaire (KCCQ) quality of life at week 52 for EF ≥50%: 4.24 vs. 2.78 (p = 0.006); EF 41-49%: 4.86 vs. 3.3 (p = 0.043; p for interaction = 0.92)

Health-related quality of life: Patients treated with empagliflozin had significant improvement in KCCQ-Clinical Summary Score (KCCQ-CSS) versus placebo (+1.03, +1.24, and +1.50 at 12, 32, and 52 weeks, respectively; p < 0.01); similar results were seen for total symptom score and overall summary score. At 12 weeks, patients in the empagliflozin arm were more likely to show meaningful improvements (≥5 points [51.6% vs. 46.5]; ≥10 points [45.0% vs. 41.8%]; ≥15 points [44.0% vs.41.3%]) and less likely to show deterioration (≥5 points [21.6% vs. 24.4%]) in KCCQ-CSS.

Pooled analysis of EMPEROR-Reduced and EMPEROR-Preserved:

Renal outcomes (profound and sustained decreases in eGFR or renal replacement therapy), total n = 9,718: 2.8% vs. 3.5% for empagliflozin vs. placebo, with significant heterogeneity between both trials (p = 0.016 for interaction).

Potassium balance (n = 9,583): Patients with hyperkalemia (potassium >5 mmol/L) were more likely to have lower EF, diabetes, and lower mean eGFR; they were also more likely to be treated with sacubitril/valsartan and an MRA. Empagliflozin reduced the composite of investigator-reported hyperkalemia or initiation of potassium binders compared with placebo (6.5% vs. 7.7%, HR 0.82, 95% CI 0.71-0.95, p = 0.01). Composite of investigator-reported hypokalemia or initiation of potassium supplements was similar compared with placebo (6.4% vs. 6.7%, p > 0.05).

Proteomics substudy (n = 1,134): Using Olink^® Explore 1536 platform, 1,283 circulating proteins were measured at baseline, week 12, and week 52. Among these, nine proteins demonstrated the largest treatment effect of empagliflozin: insulin-like growth factor-binding protein 1, transferrin receptor protein 1, carbonic anhydrase 2, erythropoietin, protein-glutamine gamma-glutamyl transferase 2, thymosin beta-10, U-type mitochondrial creatine kinase, insulin-like growth factor-binding protein 4, and adipocyte fatty acid-binding protein 4. The most common biological action of differentially expressed proteins appeared to be the promotion of autophagic flux in the heart, kidney, or endothelium, a feature of six proteins.

Effect of age: Patients were grouped based on baseline age (<65 years [n = 1,199], 65-74 years [n = 2,214], 75-79 years [n = 1,276], ≥80 years [n = 1,299]). Event rates increased with increasing age. Empagliflozin reduced primary outcomes (p for trend = 0.33), first HF hospitalization (p for trend = 0.22), and first and recurrent HF hospitalization (p for trend = 0.11) across all age groups. Effects were also similar at ≥75 years (p for interaction = 0.22) or ≥80 years (p for interaction = 0.51).

Effect of sex: 44.7% of enrolled patients were women. Compared with placebo, empagliflozin reduced the risk of CV death or HF hospitalization to a similar degree in both sexes (HR 0.81, 95% CI 0.69-0.96 for men; HR 0.75, 95% CI 0.61-0.92 for women; p for interaction = 0.54). Similar results were seen for secondary outcomes, quality of life, and physiological measures.

Effect of baseline BP: Patients were categorized into three groups based on baseline SBP: <110, 110-130, and >130 mm Hg. Among these three groups, no significant interaction in the effect of empagliflozin vs. placebo was noted for the primary endpoint (p for interaction = 0.69), first HF hospitalization (p for interaction = 0.46), or recurrent HF hospitalization (p for interaction = 0.55). SBP did not meaningfully associate with adverse events such as hypotension, volume depletion, and acute renal failure.

Concomitant diuretic use: At baseline, 20.3% were not on diuretics, 29.7% were taking furosemide (or equivalent) <40 mg, 30.5% were taking furosemide (or equivalent) 40 mg, and 19.6% were taking furosemide (or equivalent) >40 mg daily. The impact of empagliflozin vs. placebo was similar irrespective of baseline diuretic use for CV death or first HF hospitalization (p for interaction = 0.58), total HF hospitalization (p for interaction = 0.94), first HF hospitalization (p for interaction = 0.68), CV death (p for interaction = 0.31), all-cause mortality (p for interaction = 0.94), and the composite kidney endpoint (p for interaction = N/A). Empagliflozin was associated with a higher incidence of volume depletion events in the diuretic group (7.5 vs 5.6 events per 100 patient-years; HR, 1.34; 95% CI, 1.13-1.59) but not the nondiuretic group (4.3 vs 4.1 events per 100 patient-years; HR, 1.06; 95% CI, 0.70-1.61; p for interaction = 0.32). Among patients who were not taking diuretics at baseline, empagliflozin was associated with a lesser likelihood of diuretic initiation (HR, 0.73; 95% CI, 0.59-0.90; p = 0.004) compared with placebo. In those who were treated with diuretics at baseline, empagliflozin use was associated with a significantly greater probability of diuretic discontinuation (HR, 1.43; 95% CI, 1.15-1.78; p = 0.002) and de-escalation (HR, 1.15; 95% CI, 1.02-1.30; p = 0.02).

Blinded withdrawal of long-term randomized patients: At the end of the trials, 6,799 patients (placebo 3,381, empagliflozin 3,418) were prospectively withdrawn from treatment in a blinded manner, and, of these, 3,981 patients (placebo 2,020, empagliflozin 1,961) underwent prespecified in-person assessments after 30 days off treatment. From 90 days from the start of closeout to the end of double-blind treatment, the annualized risk of CV death or HF hospitalization was lower in empagliflozin-treated patients than in placebo-treated patients (10.7 vs. 13.5 events per 100 patient-years, respectively; HR 0.76 [95% CI 0.60–0.96]). When the study drugs were withdrawn for ~30 days, the annualized risk of CV death or HF hospitalization increased in patients withdrawn from empagliflozin but not in those withdrawn from placebo (17.0 vs. 14.1 events per 100 patient-years for empagliflozin and placebo, respectively). After withdrawal, the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score (KCCQ-CSS) declined by 1.6 in patients withdrawn from empagliflozin versus placebo (p < 0.0001).