Table 3.

Effects of empagliflozin on cardiovascular composite outcome (bioimpedance substudy cohort)

Outcome	Empagliflozin		Placebo		Hazard Ratio	95% CI	P Value
	No. of Participants / Total	%	No. of Participants / Total	%
Key secondary assessment
Death from heart failure, hospitalization for heart failure, development of new moderate or severe “Fluid Overload”	35/332	10.5	38/328	11.6	0.91	(0.57 to 1.45)	0.69
Death from heart failure	0/332	0.0	0/328	0.0	—	—
Hospitalization for heart failure	11/332	3.3	16/328	4.9	0.67	(0.31 to 1.46)
Development of new moderate “Fluid Overload”a	18/232	7.8	25/247	10.1	0.68	(0.37 to 1.26)
Development of new severe “Fluid Overload”b	8/302	2.6	4/303	1.3	1.96	(0.57 to 6.71)
Tertiary assessment
Regression of “Fluid Overload”c	46/84	54.8	35/72	48.6	1.33	(0.82 to 2.18)

All analyses use a time-to-first-event approach. Cox proportional hazards models include adjustment for the covariates used in the minimization algorithm: age, sex, diabetes status, eGFR, and urinary albumin-creatinine ratio. Results were consistent in post hoc sensitivity analyses additionally adjusted for use of any diuretic or loop diuretics at baseline (hazard ratios [95% CIs] 0.89 [0.56 to 1.42] and 0.92 [0.58 to 1.47]; respectively). CI, confidence interval.

^aRequires randomization value of relative “Fluid Overload” ≤7% and follow-up value >7%, ≤15%.

^bRequires randomization value of relative “Fluid Overload” ≤15% and follow-up value >15%.

^cRequires randomization value consistent with moderate or severe relative “Fluid Overload” and regression to any lower hydration category at any follow-up (limited to first event). All 660 participants were included in the composite outcome analysis because all participants were at risk of the clinical components of the composite. In the full-trial cohort, there were 88 (2.7%) first hospitalizations for heart failure in the empagliflozin group versus 107 (3.2%) in the placebo group: hazard ratio 0.80, 95% CI, 0.60 to 1.06.