Torsemide vs Furosemide Among Patients With New-Onset vs Worsening Chronic Heart Failure: A Substudy of the TRANSFORM-HF Randomized Clinical Trial

Selim R Krim; Senthil Anand; Stephen J Greene; Anqi Chen; Daniel Wojdyla; Juan Vilaro; Herbert Haught; John M Herre; Eric L Eisenstein; Kevin J Anstrom; Bertram Pitt; Eric J Velazquez; Robert J Mentz

doi:10.1001/jamacardio.2023.4776

. 2023 Nov 13;9(2):182–188. doi: 10.1001/jamacardio.2023.4776

Torsemide vs Furosemide Among Patients With New-Onset vs Worsening Chronic Heart Failure

A Substudy of the TRANSFORM-HF Randomized Clinical Trial

Selim R Krim ^1,^2,^✉, Senthil Anand ¹, Stephen J Greene ^3,⁴, Anqi Chen ³, Daniel Wojdyla ³, Juan Vilaro ⁵, Herbert Haught ⁶, John M Herre ⁷, Eric L Eisenstein ³, Kevin J Anstrom ⁸, Bertram Pitt ⁹, Eric J Velazquez ¹⁰, Robert J Mentz ^3,⁴

¹Section of Cardiomyopathy & Heart Transplantation, John Ochsner Heart and Vascular Institute, Ochsner Clinic Foundation, New Orleans, Louisiana

²The University of Queensland School of Medicine, Ochsner Clinical School, New Orleans, Louisiana

³Duke Clinical Research Institute, Durham, North Carolina

⁴Department of Medicine, Division of Cardiology, Duke University School of Medicine, Durham, North Carolina

⁵University of Florida, Gainsville

⁶Huntsville Hospital, Huntsville, Alabama

⁷Sentara Norfolk General Hospital, Norfolk, Virginia

⁸Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill

⁹Division of Cardiology, Department of Medicine, University of Michigan, Ann Arbor

¹⁰Department of Internal Medicine, Section of Cardiovascular Medicine, Yale School of Medicine, New Haven, Connecticut

Accepted for Publication: October 23, 2023.

Published Online: November 13, 2023. doi:10.1001/jamacardio.2023.4776

^✉

Corresponding Author: Selim R. Krim, MD, Section of Cardiomyopathy & Heart Transplantation, John Ochsner Heart and Vascular Institute, Ochsner Clinic Foundation, 1514 Jefferson Hwy, New Orleans, LA 70121 (selim.krim@ochsner.org).

Author Contributions: Dr Krim had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Krim, Anand, Greene, Anstrom, Velazquez, Mentz.

Acquisition, analysis, or interpretation of data: Krim, Anand, Greene, Chen, Wojdyla, Vilaro, Haught, Herre, Eisenstein, Anstrom, Velazquez, Mentz.

Drafting of the manuscript: Krim, Anand.

Critical review of the manuscript for important intellectual content: Krim, Anand, Greene, Chen, Wojdyla, Vilaro, Haught, Herre, Eisenstein, Anstrom, Velazquez, Mentz.

Statistical analysis: Chen, Wojdyla, Anstrom.

Obtained funding: Eisenstein, Velazquez, Mentz.

Administrative, technical, or material support: Herre, Velazquez.

Supervision: Krim, Greene, Haught, Anstrom, Velazquez, Mentz.

Conflict of Interest Disclosures: Dr Krim reported receiving research support from Abbott and CareDx. Dr Greene reported receiving research support from the Duke University Department of Medicine Chair’s Research Award, American Heart Association, National Heart Lung and Blood Institute, Amgen, AstraZeneca, Bristol Myers Squibb, Cytokinetics, Merck, Novartis, Pfizer, and Sanofi; serving on advisory boards for Amgen, AstraZeneca, Boehringer Ingelheim/Lilly, Bristol Myers Squibb, Cytokinetics, Roche Diagnostics, and Sanofi; serving as a consultant for Amgen, Bayer, Bristol Myers Squibb, Boehringer Ingelheim/Lilly, Corteria, CSL Vifor, Lexicon, Merck, PharmaIN, Roche Diagnostics, Sanofi, scPharmaceuticals, Tricog Health, Urovant Pharmaceuticals; and receiving speaker fees from Bayer, Boehringer Ingelheim, Cytokinetics, Lexicon, and Roche Diagnostics. Dr Herre reported serving as a consultant for Medtronic, nonexecutive chair, LifeNet Health (a not-for-profit Organ Procurement Organization and tissue bank). Dr Eisenstein reported receiving grants from the National Heart, Lung, and Blood Institute during the conduct of the study. Dr Anstrom reported receiving research support from Merck, Bayer, Pfizer, the National Institutes of Health, and Patient-Centered Oucomes Research Institute. Dr Pitt reported receiving personal fees from Bayer, scPharmaceuticals, Sqinnovation, Boehringer Ingelheim, Vifor, AstraZeneca, G3 Pharmaceuticals, Sarfez, Cereno Scientific, Phasebio, KPB Biosciences, Protonintel, Brainstorm Medical; having a patent for US 9931413 issued, site-specific delivery of eplerenone to the myocardium, and having a patent for US 63/045,783 pending, histone modulating agents for the prevention and protection from organ injury. Dr Velazquez reported receiving grants from the National Heart, Lung, and Blood Institute to Yale University as principal investigator of the clinical coordinating center during the conduct of the study and personal fees from Novartis and consultant fees from Abiomed to Yale University and from Baim Institute to Yale University outside the submitted work. Dr Mentz reported receiving research support and honoraria from Abbott outside the submitted work. No other disclosures were reported.

Funding/Support: This work was supported with grants U01-HL125478 and U01-HL125511 from the National Heart, Lung, and Blood Institute and ancillary study grants R01HL148354-04 and R01HL154768-02 from the National Heart, Lung, and Blood Institute.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Meeting Presentation: This paper was presented at the American Heart Association Scientific Sessions 2023; November 13, 2023; Philadelphia, Pennsylvania.

Data Sharing Statement: See Supplement 2.

^✉

Corresponding author.

PMCID: PMC10644243 PMID: 37955908

This post hoc analysis of the Torsemide Comparison With Furosemide for Management of Heart Failure (TRANSFORM-HF) randomized clinical trial investigates if torsemide improves mortality and hospitalization outcomes at 12 months compared with furosemide in either de novo heart failure or worsening chronic heart failure.

Key Points

Question

Does torsemide improve mortality and hospitalization outcomes at 12 months compared with furosemide in either de novo heart failure (HF) or worsening chronic HF (WHF)?

Findings

In this post hoc analysis of the Torsemide Comparison With Furosemide for Management of Heart Failure (TRANSFORM-HF) randomized clinical trial including 2858 patients, patients with de novo HF had a significantly lower all-cause mortality at 12 months when compared with WHF. No significant association of discharge diuretic type with outcomes was noted in either de novo HF or WHF.

Meaning

Among patients discharged after hospitalization for HF, de novo HF was associated with better outcomes when compared with WHF. Regardless of HF type, there was no significant difference in outcomes with a loop diuretic strategy of torsemide vs furosemide.

Abstract

Importance

Differences in clinical profiles, outcomes, and diuretic treatment effects may exist between patients with de novo heart failure (HF) and worsening chronic HF (WHF).

Objectives

To compare clinical characteristics and treatment outcomes of torsemide vs furosemide in patients hospitalized with de novo HF vs WHF.

Design, Setting, and Participants

All patients with a documented ejection fraction who were randomized in the Torsemide Comparison With Furosemide for Management of Heart Failure (TRANSFORM-HF) trial, conducted from June 18 through March 2022, were included in this post hoc analysis. Study data were analyzed March to May 2023.

Exposure

Patients were categorized by HF type and further divided by loop diuretic strategy.

Main Outcomes and Measures

End points included all-cause mortality and hospitalization outcomes over 12 months, as well as change from baseline in the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score (KCCQ-CSS).

Results

Among 2858 patients (mean [SD] age, 64.5 [14.0] years; 1803 male [63.1%]), 838 patients (29.3%) had de novo HF, and 2020 patients (70.7%) had WHF. Patients with de novo HF were younger (mean [SD] age, 60.6 [14.5] years vs 66.1 [13.5] years), had a higher glomerular filtration rate (mean [SD], 68.6 [24.9] vs 57.0 [24.0]), lower levels of natriuretic peptides (median [IQR], brain-type natriuretic peptide, 855.0 [423.0-1555.0] pg/mL vs 1022.0 [500.0-1927.0] pg/mL), and tended to be discharged on lower doses of loop diuretic (mean [SD], 50.3 [46.2] mg vs 63.8 [52.4] mg). De novo HF was associated with lower all-cause mortality at 12 months (de novo, 65 of 838 [9.1%] vs WHF, 408 of 2020 [25.4%]; adjusted hazard ratio [aHR], 0.50; 95% CI, 0.38-0.66; P < .001). Similarly, lower all-cause first rehospitalization at 12 months and greater improvement from baseline in KCCQ-CSS at 12 months were noted among patients with de novo HF (median [IQR]: de novo, 29.94 [27.35-32.54] vs WHF, 23.68 [21.62-25.74]; adjusted estimated difference in means: 6.26; 95% CI, 3.72-8.81; P < .001). There was no significant difference in mortality with torsemide vs furosemide in either de novo (No. of events [rate per 100 patient-years]: torsemide, 27 [7.4%] vs furosemide, 38 [10.9%]; aHR, 0.70; 95% CI, 0.40-1.14; P = .15) or WHF (torsemide 212 [26.8%] vs furosemide, 196 [24.0%]; aHR, 1.08; 95% CI, 0.89-1.32; P = .42; P for interaction = .10), In addition, no significant differences in hospitalizations, first all-cause hospitalization, or total hospitalizations at 12 months were noted with a strategy of torsemide vs furosemide in either de novo HF or WHF.

Conclusions and Relevance

Among patients discharged after hospitalization for HF, de novo HF was associated with better clinical and patient-reported outcomes when compared with WHF. Regardless of HF type, there was no significant difference between torsemide and furosemide with respect to 12-month clinical or patient-reported outcomes.

Introduction

Although decongestion with intravenous diuretics remains a mainstay of therapy for patients hospitalized for heart failure (HF),^1,2 significant differences in clinical profiles and outcomes exist between patients with de novo HF and worsening chronic HF (WHF). Although furosemide remains the dominant loop diuretic used in clinical practice, some patients are subsequently switched to torsemide as HF disease severity progresses, including patients who exhibit diuretic resistance with comorbid kidney disease. However, it remains unclear if an approach of early up-front use of torsemide starting at the time of HF diagnosis may improve outcomes. In this context, our study aimed to describe differences in patient characteristics and assess for a differential treatment outcome between torsemide and furosemide in de novo HF vs WHF.

Methods

Study Design

The trial design and results of the Torsemide Comparison With Furosemide for Management of Heart Failure (TRANSFORM-HF) study have been previously published.^3,4 Briefly, this open-label, pragmatic, randomized clinical trial, which was conducted from June 2018 through March 2022 with follow-up through 30 months for death and 12 months for hospitalizations, recruited 2859 participants hospitalized with HF (irrespective of ejection fraction [EF]) at 60 hospitals in the US and assessed the effect of 2 diuretic strategies (torsemide vs furosemide) on outcomes at 12 months.^3,4 The present post hoc analysis aimed to (1) characterize the patient profile and outcomes of patients with de novo HF vs WHF and (2) estimate the randomized treatment outcome (torsemide vs furosemide) associated with de novo HF vs WHF status, overall and stratified by EF category (EF ≤40%, EF >40%). The trial was approved by the Duke University institutional review board (IRB), as well as a central IRB or local site IRBs, and all patients provided written, informed consent before enrollment. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.

Study Population and End Points

All patients enrolled and successfully randomized in the TRANSFORM-HF trial⁴ were included in this analysis with the exception of patients with undocumented HF type (de novo vs WHF) and unknown EF. Patients self-identified with the following race and ethnicity categories: American Indian or Alaska Native, Asian, Black or African American, Hispanic or Latino, multiracial, Native Hawaiian or Other Pacific Islander, White, or other race and ethnicity. The inclusion of race and ethnicity data was aligned with National Institutes of Health guidance. The selected primary end point was all-cause mortality over 12 months. Secondary end points included all-cause mortality and/or first rehospitalization over 12 months, time to first all-cause rehospitalization over 12 months, total hospitalizations over 12 months, and Kansas City Cardiomyopathy Questionnaire Clinical Summary Score (KCCQ-CSS) change from baseline over 12 months.

Statistical Analysis

Patient characteristics were compared by HF type (de novo HF vs WHF). Differences between groups were assessed using the χ²/Fisher exact test for categorical comparisons as appropriate, and the Wilcoxon/Kruskal-Wallis test for comparison of continuous variables.

For the time-to-event analysis on the primary end point (all-cause mortality) and the secondary end points of all-cause death/hospitalization, Cox proportional hazards regression models were used to compare de novo vs WHF. For the secondary end point of first rehospitalization over 12 months, a Cox proportional hazards regression model was used with censoring for death. The proportional hazards assumption was confirmed using weighted Schoenfeld residuals. For the secondary end point of total hospitalizations over 12 months, relative risk was determined from a negative binomial regression of the frequency of rehospitalizations.

To compare de novo HF vs WHF, models were adjusted for prespecified covariates including HF type (de novo, WHF), age, sex, race, EF category (≤40%, 41%-49%, ≥50%, unknown), systolic blood pressure, estimated glomerular filtration rate (eGFR), hemoglobin level, body mass index, diabetes, atrial fibrillation/flutter, ischemic etiology, and chronic lung disease. For adjusted model in analyses stratified by EF levels (≤40%, >40%), the same covariates were used, with continuous baseline EF as an additional covariate. To test the interaction effect between HF types and EF category among patients with known EF level, an interaction term between baseline EF level (≤40%, >40%) and HF type was tested separately for each end point, and the model included HF type (de novo, WHF), age, sex, race, baseline EF level (≤40%, >40%), systolic blood pressure, eGFR, hemoglobin level, body mass index, diabetes, atrial fibrillation/flutter, ischemic etiology, chronic lung disease, and the interaction term.

In addition, unadjusted Kaplan-Meier curves for all-cause mortality comparing de novo and WHF were performed for the overall cohort and separately stratified by EF levels (≤40%, >40%). Unadjusted cumulative incidence curves of all-cause hospitalization comparing de novo and worsening HF, accounting for death as competing risk, were presented for the overall cohort, and separately stratified by EF levels (≤40%, >40%).

The KCCQ-CSS end point was analyzed using a linear mixed model with an unstructured covariance matrix. Repeated measures were the KCCQ-CSS score change from baseline by visit.

To estimate the randomized diuretic treatment association according to HF type with all outcomes, a model was produced for each outcome first for the overall patients, then separately for each EF category (≤40% and >40%). For the overall patient model, covariates included diuretic treatment (torsemide, furosemide), HF type, age, sex, and baseline EF categories (≤40%, 41%-49%, ≥50%, unknown). For each stratified analysis by EF level, the covariate was diuretic treatment, HF type, age, sex, and continuous baseline EF. For both overall and stratified analysis, the interaction P value between diuretic treatment and HF type was estimated separately. For the primary outcome, a 2-sided P value < .05 was considered statistically significant. For all other analyses, including secondary analyses and subgroup analyses, a P value < .005 was considered statistically significant to improve the reproducibility of study results. Study data were analyzed March to May 2023. All analyses were performed with the use of SAS software, version 9.4 (SAS Institute).

Results

Patient Characteristics

Among 2858 patients (mean [SD] age, 64.5 [14.0] years; 1803 male [63.1%]; 1055 female [36.9%]), 838 patients (29.3%) had de novo HF, and 2020 patients (70.7%) had WHF (Table 1). Patients self-identified with the following race and ethnicity categories: 12 American Indian or Alaska Native (0.4%), 63 Asian (2.2%), 968 Black or African American (33.9%), 154 Hispanic or Latino (5.4%), 43 multiracial (1.5%), 20 Native Hawaiian or Other Pacific Islander (0.7%), 1668 White (58.5%), and 79 other race or ethnicity (2.8%). Patients with de novo HF were younger (mean [SD] age, 60.6 [14.5] years vs 66.1 [13.5] years), had a higher glomerular filtration rate (mean [SD], 68.6 [24.9] vs 57.0 [24.0]), lower levels of natriuretic peptides (median [IQR], brain-type natriuretic peptide, 855.0 [423.0-1555.0] pg/mL vs 1022.0 [500.0-1927.0] pg/mL; to convert brain-type natriuretic peptide to nanograms per liter, multiply by 1), and tended to be discharged on lower doses of loop diuretic (mean [SD], 50.3 [46.2] mg vs 63.8 [52.4] mg). No significant differences in sex, race, or blood pressure were found between the 2 groups. Additionally, both groups showed a similar distribution across EF categories. Patients with WHF were more likely than patients with de novo HF to be taking an oral diuretic before the index hospital admission (torsemide: 249 of 2020 [14.4%] vs 9 of 838 [4.7%]; bumetanide: 123 of 2020 [7.1%] vs 6 of 838 [3.1%]), with furosemide being the most commonly prescribed diuretic (furosemide: 1532 of 2858 [79.8%] vs torsemide: 258 of 2858 [13.4%] and bumetanide: 129 of 2858 [6.7%]). Moreover, higher discharge diuretic dose was noted among patients with WHF compared with patients with de novo HF (mean [SD] total daily dose, 63.8 [52.4] mg vs 50.3 [46.2] mg) (Table 1 and eTables 1 and 2 in Supplement 1).

Table 1. Baseline Characteristics of Study Participants (Demographics, Laboratory, and Vital Signs, Heart Failure Characteristics and Medical History) by Heart Failure (HF) Type.

Characteristic	All patients (N = 2858)	De novo (n = 838)	WHF (n = 2020)	P value
Age, y
Mean (SD)	64.5 (14.0)	60.6 (14.5)	66.1 (13.5)	<.001
Median (IQR)	65.0 (56.0-75.0)	61.0 (51.0-71.0)	67.0 (58.0-76.0)	<.001
Sex, No. (%)
Female	1055 (36.9)	309 (36.9)	746 (36.9)	.98
Male	1803 (63.1)	529 (63.1)	1274 (63.1)	.98
Race, No. (%)
American Indian or Alaska Native	12 (0.4)	3 (0.4)	9 (0.4)	<.001
Asian	63 (2.2)	27 (3.2)	36 (1.8)
Black or African American	968 (33.9)	240 (28.7)	728 (36.1)
Multiracial	43 (1.5)	10 (1.2)	33 (1.6)
Native Hawaiian or Other Pacific Islander	20 (0.7)	6 (0.7)	14 (0.7)
White	1668 (58.5)	516 (61.6)	1152 (57.1)
Other	79 (2.8)	35 (4.2)	44 (2.2)
Ethnicity, No. (%)
Hispanic or Latino	154 (5.4)	48 (5.7)	106 (5.3)	.60
Laboratory values, mean (SD)
Sodium, mEq/L	137.8 (3.7)	137.8 (3.6)	137.8 (3.7)	.80
Potassium, mEq/L	4.04 (0.46)	4.01 (0.45)	4.05 (0.46)	.03
Blood urea nitrogen, mg/dL	33.1 (40.1)	29.2 (46.0)	34.8 (37.3)	<.001
Creatinine, mg/dL	1.403 (0.641)	1.263 (0.627)	1.461 (0.638)	<.001
Estimated glomerular filtration rate^a	60.4 (24.9)	68.6 (24.9)	57.0 (24.0)	<.001
N-terminal pro–brain-type natriuretic peptide, pg/mL
No.	1376	393	983	.01
Median (IQR)	3912.5 (1937.0-8353.5)	3569.0 (1855.0-6746.0)	4136.0 (1954.0-8808.0)	.01
Brain-type natriuretic peptide, pg/mL
No.	1381	411	970	.004
Median (IQR)	964.7 (474.0-1809.0)	855.0 (423.0-1555.0)	1022.0 (500.0-1927.0)	.004
Hemoglobin, mean (SD), g/L	12.10 (4.72)	12.58 (2.48)	11.90 (5.37)	<.001
Vital signs
Systolic blood pressure, mean (SD), mm Hg	118.8 (19.8)	119.6 (19.8)	118.4 (19.9)	.18
Diastolic blood pressure, mean (SD), mm Hg	69.7 (12.9)	71.7 (12.9)	68.9 (12.7)	<.001
Heart rate, mean (SD), bpm	80.6 (15.8)	83.8 (16.1)	79.3 (15.5)	<.001
BMI, mean (SD)	32.18 (9.49)	32.45 (9.21)	32.07 (9.60)	.18
Heart failure characteristics
Ejection fraction rate, No. (%)
≤40	1835 (64.2)	558 (66.6)	1277 (63.2)	.09
>40	799 (28.0)	227 (27.1)	572 (28.3)
Unknown	224 (7.8)	53 (6.3)	171 (8.5)
Duration of diagnosis prior to hospitalization, No. (%)
≤ 30 d	119 (5.9)	0	119 (5.9)
> 30 d and ≤12 mo	345 (17.1)	0	345 (17.1)
> 12 mo	1365 (67.6)	0	1365 (67.6)
Unknown	191 (9.5)	0	191 (9.5)
NYHA class at hospital admission, No. (%)
I	35 (1.2)	7 (0.8)	28 (1.4)	.001
II	284 (10.0)	88 (10.5)	196 (9.7)
III	1190 (41.7)	320 (38.3)	870 (43.2)
IV	404 (14.2)	102 (12.2)	302 (15.0)
Unknown	939 (32.9)	319 (38.2)	620 (30.8)
NYHA class at randomization, No. (%)
I	70 (2.4)	15 (1.8)	55 (2.7)	<.001
II	651 (22.8)	206 (24.6)	445 (22.0)
III	977 (34.2)	246 (29.4)	731 (36.2)
IV	196 (6.9)	53 (6.3)	143 (7.1)
Unknown	964 (33.7)	318 (37.9)	646 (32.0)
KCCQ Baseline Clinical Summary Score
Mean (SD)	43.18 (23.23)	46.27 (23.43)	41.91 (23.03)	<.001
Median (IQR)	41.1 (25.0-59.8)	45.3 (28.1-62.8)	39.6 (24.0-58.3)	<.001
Ischemic etiology, No. (%)	808 (28.3)	162 (19.3)	646 (32.0)	<.001
Chronic kidney disease, No. (%)	1008 (35.3)	144 (17.2)	864 (42.8)	<.001
Diabetes, No. (%)	1363 (47.7)	319 (38.1)	1044 (51.7)	<.001
Hypertension, No. (%)	2329 (81.7)	609 (72.9)	1720 (85.3)	<.001
Atrial fibrillation or atrial flutter, No. (%)	1274 (44.9)	254 (30.5)	1020 (50.8)	<.001
Chronic lung disease (including COPD), No. (%)	675 (23.8)	112 (13.5)	563 (28.0)	<.001
Loop diuretics
Oral loop diuretic before index hospital admission (most recent), No. (%)
No	939 (32.9)	645 (77.0)	294 (14.6)	<.001
Yes	1919 (67.1)	193 (23.0)	1726 (85.4)	<.001
Total daily dose prescribed (furosemide equivalents), mg
No.	1901	190	1711	<.001
Mean (SD)	66.1 (61.5)	43.8 (28.8)	68.6 (63.6)	<.001
Specific loop diuretic before index admission (among those on a loop diuretic), No. (%)^b
Furosemide	1532 (79.8)	178 (92.2)	1354 (78.4)
Torsemide	258 (13.4)	9 (4.7)	249 (14.4)
Bumetanide	129 (6.7)	6 (3.1)	123 (7.1)
Ethacrynic acid	0	0	0
Discharge, No. (%)
Prescribed assigned loop diuretic^a	2490 (90.4)	735 (91.3)	1755 (90.0)	.05
Total daily dose prescribed (furosemide equivalents), mean (SD), mg^b	79.3 (63.3)	64.9 (51.7)	85.4 (66.6)	<.001
Total daily dose prescribed (reported), mean (SD), mg	59.8 (51.0)	50.3 (46.2)	63.8 (52.4)	<.001
Baseline medications taken at the time of randomization, No. (%)
ACE or ARB	1243 (43.5)	411 (49.0)	832 (41.2)	<.001
ARNi^c	536 (18.8)	142 (16.9)	394 (19.5)	.11
ACE or ARB or ARNi	1717 (60.1)	533 (63.6)	1184 (58.6)	.01
Aldosterone antagonist/MRA	1021 (35.7)	280 (33.4)	741 (36.7)	.10
β-Blocker	2245 (78.6)	613 (73.2)	1632 (80.8)	<.001
SGLT-2 inhibitor	170 (6.2)	45 (5.5)	125 (6.4)	.35
Long-acting nitrate	391 (13.7)	78 (9.3)	313 (15.6)	<.001
Thiazide diuretic	126 (4.4)	35 (4.2)	91 (4.5)	.69
Digoxin	168 (5.9)	33 (3.9)	135 (6.7)	.005
Hydralazine	432 (15.1)	104 (12.4)	328 (16.2)	.01
Ivabradine	8 (0.3)	1 (0.1)	7 (0.3)	.45

Open in a new tab

Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; ARNi, angiotensin receptor-neprilysin inhibitor; bpm, beats per minute; COPD, chronic obstructive pulmonary disease; KCCQ, Kansas City Cardiomyopathy Questionnaire; MRA, mineralocorticoid receptor antagonist; NYHA, New York Heart Association; SGLT-2, sodium-glucose transport protein 2.

SI conversion factor: To convert brain-type natriuretic peptide to nanograms per liter, multiply by 1; to convert creatinine to micromoles per liter, multiply by 88.4; to convert hemoglobin to grams per deciliter, divide by 10; to convert potassium to millimoles per liter, multiply by 1; to convert sodium to millimoles per liter, multiply by 1.

^{^a}

Denominator is the patients with known prescription status of the assigned loop diuretic at each visit (ie, excludes patients who died by the time of the visit or who had unknown prescription status at the visit).

^{^b}

Loop diuretic doses were converted to furosemide equivalents with 1-mg bumetanide = 20-mg torsemide = 50-mg ethacrynic acid = 40-mg furosemide for oral diuretics; and 1-mg bumetanide = 50-mg ethacrynic acid = 20-mg furosemide for intravenous diuretics.

^{^c}

ARNi was reported as being taken (yes) or missing. For calculating the percentage of patients taking this medication at baseline, if it was not reported as yes, it was assumed to be no.

Outcomes for Patients With De Novo HF vs WHF

De novo HF was associated with lower all-cause 12-month mortality (de novo, 65 of 838 [9.1%] vs WHF, 408 of 2020 [25.4%]; adjusted hazard ratio [aHR], 0.50; 95% CI, 0.38-0.66; P < .001) (eTable 3 and eFigure 1 in Supplement 1). After multivariate adjustments and stratification by EF, this finding persisted among patients with EF of 40% or less but not in patients with EF greater than 40% (eTable 3, eFigures 2 and 3 in Supplement 1).

Findings were generally similar in analyses of composite all-cause mortality or rehospitalization over 12 months, all-cause hospitalization over 12 months, and total hospitalizations over 12 months (eTables 3 and 4 and eFigures 4, 5, and 6 in Supplement 1).

Similarly, lower all-cause first rehospitalization at 12 months and greater improvement from baseline in KCCQ-CSS at 12 months were noted among patients with de novo HF (median [IQR]: de novo, 29.94 [27.35-32.54] vs WHF, 23.68 [21.62-25.74]; adjusted estimated difference in means: 6.26; 95% CI, 3.72-8.81; P < .001). After adjustment and EF stratification, this finding remained present only among patients with EF of 40% or less (eTable 5 in Supplement 1).

Outcome of Torsemide vs Furosemide in De Novo HF vs WHF

In comparing a treatment strategy of torsemide vs furosemide, there was no statistically significant difference in 12-month mortality among de novo HF (No. of events [rate per 100 patient-years]: torsemide, 27 [7.4%] vs furosemide, 38 [10.9%]; aHR, 0.70; 95% CI, 0.40-1.14; P = .15) or WHF (torsemide 212 [26.8%] vs furosemide, 196 [24.0%]; aHR, 1.08; 95% CI, 0.89-1.32; P = .42; P for interaction = .10) (Table 2 and Figure). Moreover, when stratified by EF, there was an associated trend in reduced mortality at 12 months among the torsemide group in patients with de novo HF with EF of 40% or less, but this finding was not seen among patients with WHF (Table 2 and eFigures 7 and 8 in Supplement 1).

Table 2. Time-to-Event Outcomes Comparison Between Randomized Diuretic Treatment Groups Stratified by De Novo vs Worsening Chronic Heart Failure (WHF).

Outcomes	De novo				WHF				P for interaction^c
Outcomes	Torsemide events (rate per 100 PY)^a	Furosemide events (rate per 100 PY)^a	HR/RR (T vs F)^b	P value	Torsemide rate (rate per 100 PY)^a	Furosemide rate (rate per 100 PY)^a	HR/RR (T vs F)^b	P value	P for interaction^c
All-cause mortality over 12 mo ^a ^, ^b
Overall	27 (7.4)	38 (10.9)	0.70 (0.4-1.14)	.15	212 (26.8)	196 (24.0)	1.08 (0.89-1.32)	.42	.10
EF level, %
≤40	13 (5.4)	25 (10.9)	0.50 (0.26-0.99)	.04	140 (27.7)	122 (24.4)	1.10 (0.86-1.41)	.45	.03
>40	11 (10.7)	8 (8.2)	1.22 (0.49-3.07)	.67	53 (23.0)	56 (23.6)	0.93 (0.63-1.36)	.71	.59
All-cause mortality or rehospitalization over 12 mo
Overall	137 (61.4)	156 (73)	0.86 (0.68-1.08)	.20	535 (114.8)	545 (121.9)	0.94 (0.83-1.06)	.33	.49
EF level, %
≤40	84 (55.2)	89 (60.9)	0.93 (0.69-1.25)	.63	345 (115.7)	333(121.9)	0.94 (0.81-1.10)	.47	.92
>40	45 (76.6)	50 (93.1)	0.87 (0.58-1.30)	.49	151 (110.9)	160 (124.2)	0.90 (0.72-1.13)	.35	.88
All-cause first rehospitalization over 12 mo
Overall	121 (54.8)	132 (63.9)	0.88 (0.69-1.13)	.32	411 (94.5)	442 (103.9)	0.91 (0.79-1.04)	.17	.83
EF level, %
≤40	74 (49.0)	72 (51.0)	0.99 (0.72-1.8)	.97	264 (95.0)	267 (102.1)	0.93 (0.78-1.10)	.38	.71
>40	39 (67.6)	46 (88.4)	0.81 (0.53-1.25)	.35	118 (92.0)	134 (110.2)	0.85 (0.66-1.09)	.21	.85

Open in a new tab

Abbreviations: EF, ejection fraction; F, furosemide; HR, hazard ratio; PY, patient-years; RR, rate ratio; T, torsemide.

^{^a}

For each cell, for all-cause mortality and all-cause mortality or rehospitalization over 12 months, the numbers displayed are number of events (Kaplan-Meier failure rate %) event rate per 100 PY of follow-up. For all-cause first rehospitalization over 12 months, the numbers displayed are number of events (cumulative incidence rate %) event rate per 100 PY of follow-up.

^{^b}

For all-cause mortality and all-cause mortality/re-hospitalization over 12 months, HR, 95% CI, and P value are based on a Cox proportional hazards regression model. For all-cause first primary rehospitalization over 12 months, HR, 95% CI, and P value is based on a Cox proportional hazards model where death is considered as competing risk and is censored at time of death. For the overall patient model, the covariates are diuretic treatment (T, F), HF type, age, sex, baseline EF categories (≤40%, 41%-49%, ≥50%, unknown). For each stratified analysis by EF level, the covariates are diuretic treatment, HF type, age, sex, and continuous baseline EF.

^{^c}

The interaction P values are between diuretic treatment and HF type for overall patients and each EF level.

Figure. — Forest plot of hazard ratio (HR)/rate ratio (RR) comparing randomized diuretic treatments (torsemide vs furosemide) for all end points among patients with de novo heart failure and patients with worsening chronic heart failure (WHF).

^aFirst event rates per 100 patient-years are reported for all-cause mortality, all-cause mortality or hospitalization, and all-cause hospitalization.

^bFor total hospitalizations, the total event rates per 100 patient-years are reported.

Additionally, irrespective of randomized loop diuretic, there were no significant differences in all-cause mortality or rehospitalization, first all-cause hospitalization, or total hospitalizations at 12 months, in either de novo HF or WHF (Table 2 and eTable 6 in Supplement 1). No significant difference in change in KCCQ-CSS was noted for either de novo HF or WHF (eTable 7 and eFigures 7 and 8 in Supplement 1).

Study Drug Adherence

Diuretic adherence at discharge and month 1, 6, and 12, respectively, is displayed in eTable 8 in Supplement 1. At the time of discharge, overall rates of diuretic discontinuation were low (76 of 2786 [2.8%]) with no significant differences among those with de novo HF or WHF. However, a steady increase in discontinuation of loop diuretics was noted among patients with de novo HF at month 1 (123 of 2786 [6.5%]), month 6 (142 of 2786 [9.1%]), and month 12 (164 of 2786 [12.9%]).

Discussion

In the TRANSFORM-HF trial, de novo HF status was associated with better outcomes as compared with WHF. However, irrespective of de novo vs WHF status, results of this post hoc analysis suggest that there was no significant difference between torsemide and furosemide with respect to 12-month clinical or patient-reported outcomes.

Consistent with previous data,^5,6,7 patients with de novo HF were slightly younger and more likely to have a nonischemic etiology when compared with those with WHF. Yet, to our knowledge, no prior studies have explored the differential association of discharge diuretic type with outcome in de novo HF vs WHF. Although torsemide has been shown to have better absorption and bioavailability when compared with furosemide,⁸ we found no significant association between diuretic type and selected primary and secondary end points in either de novo HF or WHF. Despite the observed nominally significant reduction in mortality at 12 months among the torsemide group in patients with de novo HF and an EF of 40% or less (with significant interaction P value of .03), the low event rates among patients with de novo HF and absence of a similar trend with other clinical end points makes this isolated finding likely related to chance and, therefore, not clinically relevant.

Limitations

Limitations of this study should be noted. Sample size, crossover rates, and diuretic discontinuation during the follow-up period may all have impacted our results. Despite statistical adjustment, residual or unmeasured confounding may have influenced our results.

Conclusions

In this post hoc analysis of the TRANSFORM-HF randomized clinical trial, among patients discharged with acute HF, de novo HF was associated with better outcomes when compared with WHF. Regardless of de novo HF vs WHF status, there was no significant difference between torsemide and furosemide with respect to 12-month clinical or patient-reported outcomes. Importantly, the observed high mortality among the WHF population and persistence of congestion at the time of randomization merits further investigation.

Supplement 1.

eTable 1. Baseline Characteristics of Study Participants (Demographics, Laboratory, and Vital Signs, Heart Failure Characteristics, and Medical History) by HF Type, for Patients With ≤40% EF

eTable 2. Baseline Characteristics of Study Participants (Demographics, Laboratory, and Vital Signs, Heart Failure Characteristics, and Medical History) by HF Type, for Patients With >40% EF

eTable 3. Time-to-Event Outcome Comparison Between De Novo and WHF for Overall Patients and Stratified by Ejection Fraction

eTable 4. Effect of HF Type on Total Rehospitalizations Through Month 12 for Overall Patients and stratified by ejection fraction

eTable 5. Comparison of KCCQ Clinical Summary Score by HF Type for overall patients and by EF Levels–Least Square Mean by Visit

eTable 6. Treatment Effect on Total Rehospitalizations Through Month 12 by HF Type for Overall Patients and Stratified by EF Levels

eTable 7. Comparison of KCCQ Clinical Summary Score Between Treatment Diuretics Groups by HF Types–Least Square Mean by Visit for Overall Patients and Stratified by EF Levels

eTable 8. Diuretic Dose Adherence at Discharge, Month 1, 6, and 12 by Heart Failure Type and Diuretic Treatment

eFigure 1. Unadjusted Kaplan-Meier curves for time to all-cause mortality comparing De novo and worsening HF for overall patients over 12 months

eFigure 2. Unadjusted Kaplan-Meier Curves for Time to All-Cause Mortality Comparing De Novo and WHF for Patients With EF Level ≤40% Over 12 Months

eFigure 3. Unadjusted Kaplan-Meier Curves for Time to All-Cause Mortality Comparing De Novo and WHF for Patients With EF Level >40% Over 12 Months

eFigure 4. Unadjusted Cumulative Incidence Curves for All-Cause Hospitalization Over 12 Months Comparing De Novo and WHF for Overall Patients

eFigure 5. Unadjusted Cumulative Incidence Curves for All-Cause Hospitalization Over 12 Months Comparing De Novo and WHF for Patients With EF Level ≤40%

eFigure 6. Unadjusted Cumulative Incidence Curves for All-Cause Hospitalization Over 12 Months Comparing De Novo and WHF for Patients With EF Level >40%

eFigure 7. Forest Plot of HR/RR Comparing Randomized Diuretic Treatments (Torsemide vs Furosemide) for Overall Patients and Patients for All End Points Among Patient With EF Level ≤40%

eFigure 8. Forest Plot of HR/RR Comparing Randomized Diuretic Treatments (Torsemide vs Furosemide) for Overall Patients and Patients for All End Points Among Patient With EF Level >40%

Click here for additional data file.^{(1.4MB, pdf)}

Supplement 2.

Data Sharing Statement.

Click here for additional data file.^{(13KB, pdf)}

References

1.Sinnenberg L, Givertz MM. Acute heart failure. Trends Cardiovasc Med. 2020;30(2):104-112. doi: 10.1016/j.tcm.2019.03.007 [DOI] [PubMed] [Google Scholar]
2.Mebazaa A, Yilmaz MB, Levy P, et al. Recommendations on pre-hospital and early hospital management of acute heart failure: a consensus paper from the Heart Failure Association of the European Society of Cardiology, the European Society of Emergency Medicine and the Society of Academic Emergency Medicine–short version. Eur Heart J. 2015;36(30):1958-1966. doi: 10.1093/eurheartj/ehv066 [DOI] [PubMed] [Google Scholar]
3.Greene SJ, Velazquez EJ, Anstrom KJ, et al. ; TRANSFORM-HF Investigators . Pragmatic design of randomized clinical trials for heart failure: rationale and design of the TRANSFORM-HF Trial. JACC Heart Fail. 2021;9(5):325-335. doi: 10.1016/j.jchf.2021.01.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Mentz RJ, Anstrom KJ, Eisenstein EL, et al. ; TRANSFORM-HF Investigators . Effect of torsemide vs furosemide after discharge on all-cause mortality in patients hospitalized with heart failure: the TRANSFORM-HF randomized clinical trial. JAMA. 2023;329(3):214-223. doi: 10.1001/jama.2022.23924 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Greene SJ, Hernandez AF, Dunning A, et al. Hospitalization for recently diagnosed vs worsening chronic heart failure: from the ASCEND-HF Trial. J Am Coll Cardiol. 2017;69(25):3029-3039. doi: 10.1016/j.jacc.2017.04.043 [DOI] [PubMed] [Google Scholar]
6.Butt JH, Fosbøl EL, Gerds TA, et al. Readmission and death in patients admitted with new-onset vs worsening of chronic heart failure: insights from a nationwide cohort. Eur J Heart Fail. 2020;22(10):1777-1785. doi: 10.1002/ejhf.1800 [DOI] [PubMed] [Google Scholar]
7.Fonarow GC, Abraham WT, Albert NM, et al. ; OPTIMIZE-HF Investigators and Hospitals . Factors identified as precipitating hospital admissions for heart failure and clinical outcomes: findings from OPTIMIZE-HF. Arch Intern Med. 2008;168(8):847-854. doi: 10.1001/archinte.168.8.847 [DOI] [PubMed] [Google Scholar]
8.Buggey J, Mentz RJ, Pitt B, et al. A reappraisal of loop diuretic choice in heart failure patients. Am Heart J. 2015;169(3):323-333. doi: 10.1016/j.ahj.2014.12.009 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials