Principal Findings:

The primary outcome, cardiovascular death or worsening HF for dapagliflozin vs. placebo, was: 16.4% vs. 19.5% (hazard ratio [HR] 0.82, 95% confidence interval [CI] 0.73-0.92, p < 0.001).

• Cardiovascular death: 7.4% vs. 8.3% (HR 0.88, 95% CI 0.74-1.05)
• HF hospitalization or urgent visit for HF: 11.8% vs. 14.5% (HR 0.79, 95% CI 0.69-0.91)

For the primary outcome, benefit was similar among patients with or without DM2, and for categories of baseline EF (≤49%, 50-59%, ≥60%).

Secondary outcomes:

• All-cause mortality: 15.9% vs. 16.8% (HR 0.94, 95% CI 0.83-1.07, p > 0.05)
• Any amputation: 0.6% vs. 0.8%
• Any major hypoglycemic event: 0.2% vs. 0.2%

Impact of frailty: Frailty was measured using the Rockwood cumulative deficit approach. On this scale, 37.6% had class 1 frailty (not frail); 38.6% had mild-moderate frailty, and 23.8% had severe frailty. The effect of dapagliflozin on the primary endpoint from Frailty Index class 1 to 3 was: 0.85 (95% CI 0.68-1.06); 0.89 (0.74-1.08); and 0.74 (0.61-0.91), respectively (p for interaction = 0.40). Although frailer patients had worse Kansas City Cardiomyopathy Questionnaire (KCCQ) scores at baseline, the improvement with dapagliflozin was greater than in less frail patients (p for interaction = 0.021).

Impact of body mass index (BMI): Mean BMI was 29.8 kg/m². Compared with placebo, dapagliflozin reduced the risk of the primary outcome to a similar extent across BMI categories (p for interaction = 0.82). The placebo-corrected change in weight at 12 months for BMI categories was: 18.5-24.9 kg/m²: –0.88; 25-29.9 kg/m²: –0.65; 30-34.9 kg/m²: –1.42; 35.0-39.9 kg/m²: –1.17; ≥40 kg/m²: –2.50 (p for interaction = 0.002). The placebo-corrected change in KCCQ-total symptom score with dapagliflozin at 8 months for the respective BMI categories was: 0.9 vs. 2.5 vs. 1.9 vs. 2.7 vs. 8.6 points (p interaction = 0.03).

Lifetime benefit modeling: Nonparametric age-based methods were used to extrapolate potential gains in survival free from the primary endpoint from long-term use of dapagliflozin. At age 55 years, the estimated survival free from the primary endpoint was 11.8 years with dapagliflozin vs. 9.8 years with placebo (p = 0.14). At age 65 years, the estimated event-free survival was 12.1 years with dapagliflozin and 9.7 years with placebo (p = 0.002). At age 75 years, the estimated event-free survival was 10.6 years with dapagliflozin and 9.4 years with placebo (p = 0.063).

Effect of N-terminal pro–B-type natriuretic peptide (NT-proBNP): The median baseline concentration of NT-proBNP was 716 ng/L and 1,399 ng/L for non–atrial fibrillation/flutter and atrial fibrillation/flutter, respectively. Higher NT-proBNP levels were linearly associated with a greater risk of the primary outcome. The clinical benefit of dapagliflozin was present for the primary outcome (p value for interaction = 0.40 by quartiles) and for HF hospitalizations (p for interaction = 0.86) irrespective of baseline NT-proBNP concentration.

Impact of baseline systolic blood pressure (SBP): Dapagliflozin reduced SBP by 1.8 (95% CI 1.1-2.5) mm Hg compared with placebo at 1 month. The treatment effect of dapagliflozin on the primary outcome and KCCQ total symptom score was consistent across SBP categories (<120, 120-129, 130-139, ≥140 mm Hg; interaction p = 0.15 and p = 0.98, respectively).

Impact of duration of HF: The number of patients in each category was as follows: 1,160 (≤6 months), 842 (>6 to 12 months), 995 (>1 to 2 years), 1,569 (>2 to 5 years), and 1,692 (>5 years). Patients with longer-duration HF were older and had more comorbidities and had worse outcomes. The benefit of dapagliflozin was consistent across HF duration category: the hazard ratio for the primary outcome in the ≤6-month group was 0.67 (95% CI 0.50-0.91); >6 to 12 months, 0.78 (0.55-1.12); >1 to 2 years, 0.81 (0.60-1.09); >2 to 5 years, 0.97 (0.77-1.22); and >5 years, 0.78 (0.64-0.96; p for interaction = 0.41).

Diuretic utilization: 10.9% were on no diuretic, 12.3% were on a non-loop diuretic, and 76.8% were on a loop diuretic at baseline. Treatment benefits of dapagliflozin on the primary composite outcome were consistent by diuretic use categories (p for interaction = 0.64) or loop diuretic dose (p for interaction = 0.57). Dapagliflozin reduced new initiation of loop diuretics by 32% (hazard ratio [HR] 0.68, 95% CI 0.55-0.84, p < 0.001) but did not influence discontinuations/disruptions (HR 0.98, 95% CI, 0.86–1.13, p = 0.83).

Deterioration in estimated glomerular filtration rate (eGFR) <25ml/min/1.73m² (pooled analysis of DELIVER and DAPA-HF): 3.2% experienced a deterioration in eGFR to <25 at least once during follow-up. The median time to deterioration in kidney function to eGFR <25ml/min/1.73m² was 121 days. The risk of the primary outcome was lower with dapagliflozin compared with placebo both among patients who did (HR 0.53, 95% CI 0.33-0.83) and did not (HR 0.78, 95% CI 0.72-0.86) experience deterioration in kidney function below eGFR 25ml/min/1.73m² (p for interaction = 0.17). Rates of drug discontinuation were similar.

Worsening of HF: Over the duration of follow-up, 72% of patients experienced no worsening HF event, while 13% had outpatient oral diuretic intensification, 1% required an urgent HF visit, 9% had a HF hospitalization, and 4% died of cardiovascular causes as a first presentation. All-cause mortality: no worsening HF: 4/100 patient-years (PY), outpatient oral diuretic intensification: 10/100 PY, urgent HF visit: 10/100 PY, HF hospitalization: 35/100 PY. Dapagliflozin relative to placebo consistently reduced the occurrence of first nonfatal worsening HF events including outpatient oral diuretic intensification (HR 0.72, 95% CI 0.64-0.82) and HF hospitalization (HR 0.77, 95% CI 0.67-0.89).

Apparent treatment-resistant hypertension: 11.5% of patients had apparent treatment-resistant hypertension at baseline (defined as BP ≥140/90 mm Hg [≥130/80 mm Hg if diabetes]) despite treatment with three antihypertensive drugs including a diuretic). The primary outcome for controlled BP vs. nonresistant hypertension group vs. apparent treatment-resistant hypertension: 8.7 vs. 8.5 vs. 9.5 per 100 patient-years. The relative benefits of dapagliflozin on the primary composite outcome were consistent irrespective of baseline BP category (HR 0.91 [95% CI 0.78–1.05] for controlled BP; HR 0.73 [95% CI 0.58–0.91] for nonresistant hypertension; HR 0.67 [95% CI 0.48–0.94] for apparent treatment-resistant HTN; p for interaction = 0.114). Similar findings were observed for individual components of the primary endpoint and across key secondary endpoints. Mean SBP reduction was 1-3 mm Hg with dapagliflozin vs. placebo.

Decline in eGFR: No decline in eGFR in 42% of patients, 26% had 0-10% decline (51% dapagliflozin, 49% placebo), 32% had decline in eGFR of >10% (61% dapagliflozin, 39% placebo). Median change in eGFR was -4 for those assigned to dapagliflozin and -1 for placebo (p < 0.001). Patients with >10% decline had a higher risk of the primary endpoint compared with those who did not (HR 1.33, 95% CI 1.10-1.62). This difference was, however, not noted among patients who were randomized to dapagliflozin (p for interaction = 0.01). No difference was noted for the primary kidney endpoint based on change in eGFR or randomization to dapagliflozin vs. placebo.

Concomitant beta-blocker use: 83% were on beta-blockers in the trial. Patients taking a beta-blocker had a lower risk for HF events, cardiovascular death, and total HF events and cardiovascular death in crude, covariate adjusted-, and propensity score-based models. The treatment effect of dapagliflozin on the primary composite was similar in patients taking beta-blockers (HR 0.82 [95% CI 0.72-0.94]) and those not taking beta-blockers (HR 0.79 [95% CI 0.61-1.03]; p for interaction = 0.85). Similarly, the benefits of dapagliflozin on worsening HF events, cardiovascular death, all-cause mortality, and total HF events and cardiovascular death did not differ by beta-blocker use (p for interaction > 0.20 for all).