Background: Sodium-glucose co-transporter-2 inhibitors are a promising new treatment option for patients with heart failure and reduced ejection fraction (HFrEF) with and without type 2 diabetes. We investigated the effects of empagliflozin on estimated extracellular volume (ECV), estimated plasma volume, and measured glomerular filtration rate (GFR) in patients with HFrEF. Methods: Empire HF Renal was an investigator-initiated, double-blind, placebo-controlled, randomised trial with recruitment from four hospitals in Denmark. Patients in New York Heart Association class I-III, with a left ventricular ejection fraction of 40% or lower, and on guideline-directed heart failure therapy were randomly assigned (1:1) to receive either oral empagliflozin 10 mg or placebo once daily for 12 weeks. The co-primary renal outcomes were the between-group difference in the change of estimated ECV, estimated plasma volume, and measured GFR from baseline to 12 weeks. The trial is registered at ClinicalTrials.gov (NCT03198585). Findings: Between June 29, 2017, and July 15, 2019, 391 patients were screened and 120 (31%) were assigned to empagliflozin or placebo; 59 (98%) in the empagliflozin group and 58 (97%) in the placebo group were included in the analyses. Significant reductions in estimated ECV, estimated plasma volume, and measured GFR with empagliflozin versus placebo were observed [estimated ECV; adjusted mean difference of change empagliflozin versus placebo -0·12 L; 95% confidence interval (CI) -0·18 to -0·05, P=0·0006. Estimated plasma volume; -7·3%; 95% CI -10·3 to -4·3, P<0·0001. Measured GFR; -7·5 mL/min; 95% CI -11·2 to -3·8, P=0·0001]. These findings were consistent across important clinical heart failure subgroups. In the empagliflozin group, five patients (8·3%) had one or more serious adverse events, versus three patients (5·0%) in the placebo group. Interpretation: In patients with HFrEF, empagliflozin reduced estimated ECV, estimated plasma volume, and measured GFR after 12 weeks. Trial Registration Number: The trial is registered with ClinicalTrials.gov (number NCT03198585) and EudraCT (number 2017-001341-27). Funding: This work was supported by the Research Council at Herlev and Gentofte University Hospital, Herlev, Denmark [institutional research grant]; the Research and Innovation Foundation of the Department of Cardiology (FUHAS, formerly FUKAP), Herlev and Gentofte University Hospital, Herlev, Denmark [institutional research grant]; the Capital Region of Denmark, Copenhagen, Denmark [grant number A6058]; the Danish Heart Foundation, Copenhagen, Denmark [grant numbers 17-R116-A7714-22076, 18 R124-A8573-22107]; the A.P. Moller Foundation for the Advancement of Medical Science, Copenhagen, Denmark [grant number 17-L-0002]. Conflict of Interest: JJ reports grants from the Research Council at Herlev and Gentofte University Hospital, Denmark, grants from the Research and Innovation Foundation of the Department of Cardiology ((FUHAS, formerly FUKAP), Herlev and Gentofte University Hospital, Denmark, and grants from the A.P. Moller Foundation for the Advancement of Medical Science, Denmark, during the conduct of the study. MO reports grants from The Steno Diabetes Center Odense, Denmark, during the conduct of the study. CK reports personal fees from scientific advisory panel and speaker fees from Boehringer Ingelheim, Merck, Sharp & Dohme, AstraZeneca, Amgen, Novartis,Novo Nordisk, and Shire, outside the submitted work. CT has nothing to disclose. IG has nothing to disclose. LK reports personal fees from speaker honorarium from Novartis, AstraZeneca, Novo Nordisk, and Boehringer Ingelheim, outside the submitted work. FG reports personal fees from Boehringer Ingelheim, during the conduct of the study; personal fees from Novartis, grants and personal fees from Pfizer, personal fees from Orion Pharma Abbott, Bayer, AstraZeneca, and Carmat, outside the submitted work. JF has nothing to disclose. MEM has nothing to disclose. ELF has nothing to disclose. NEB has nothing to disclose. JLF has nothing to disclose. LTJ has nothing to disclose. JEM reports grants and personal fees from Abiomed, personal fees from Novartis, and Orion Pharma, outside the submitted work. MS reports grant from The Capital Region of Denmark, and grants from The Danish Heart Foundation, during the conduct of the study; personal fees and non-financial support from AstraZeneca, personal fees from Novo Nordisk and Boehringer Ingelheim, outside the submitted work. Ethical Approval: The trial was approved by the local ethics committee, and all patients provided written informed consent (Danish National Committee on Health Research Ethics, number H-17010756).