Torsemide comparison with furosemide for management of heart failure (TRANSFORM-HF) trial

The torsemide comparison with furosemide formanagement of heart failure (TRANSFORM-HF) trial is a large-scale, pragmatic clinical effectiveness study comparing torsemide vs. furosemide as a treatment for heart failure (HF). The trial was event-driven and recruited inhospital patients with new or worsening HF (HF did not need to be the primary reason for admission). The primary objective of the TRANSFORM-HF trial was to evaluate whether the type of loop diuretic prescribed for oral maintenance treatment and prevention of congestion impacts on clinical outcomes. Torsemide compared with furosemide has greater and more stable bioavailability (especially when taken with food), a higher degree of protein binding and therefore a longer half-life with longer duration of action.1 In addition, it has potential vasodilatory and antifibrotic effects through inhibition of aldosterone.2 In a meta-analysis including 19 studies with over 19 000 patients, torsemide compared with furosemide use in HF was associated with a significant greater improvement in functional status lower cardiovascular mortality.3 In the TRANSFORM-HF trial, patients received 30-day, 6-month, and 12-month follow-up phone contacts for assessments of vital status, interval hospitalizations, adherence, and quality of life. The trial was registered at clinicaltrials.gov (NCT03296813). Randomized (allocation 1:1 to furosemide:torsemide) Open-label Parallel-arm (two arms) Multicentre: 60 study locations in the USA Follow-up time: up to 30 months (12 months initially planned) Total number of trial participants: 2859 (1431 randomized to torsemide; 1428 randomized to furosemide) Mean age: 65 years Men/women: 63%/37% White/black/other race: 59%/34%/7% Newly diagnosed HF: 30% Ischaemic aetiology of HF: 29% Ejection fraction: reported in 93% of the participants HF with reduced ejection fraction (HFrEF): 64% HF with mildly reduced ejection fraction: 6% HF with preserved ejection fraction (HFpEF): 23% Median N-terminal of pro-hormone B-type natriuretic peptide (NT-proBNP): ∼3900 ng/L Mean estimated glomerular filtration rate: 60 mL/min/1.73 m² Median (interquartile range) furosemide equivalent dose: 66 mg Furosemide/torsemide/bumetanide use: 79%/14%/7% 1. Patient ≥18 years of age hospitalized (≥24 h or over a change in calendar date) with worsening of chronic HF or new diagnosis of HF and either one or both of the following criteria: (a) Left ventricular EF ≤40% within 24 months prior to and including index hospitalization by any method (b) Elevated natriuretic peptide level during index hospitalization as measured by local laboratory 2. Plan for a daily outpatient oral loop diuretic regimen upon hospital discharge with anticipated need for long-term loop diuretic use. End-stage renal disease requiring renal replacement therapy History of heart transplant or actively listed for heart transplant Implanted left ventricular assist device or implant anticipated <3 months Pregnant or nursing women Malignancy or other non-cardiac condition limiting life expectancy to <12 months Known hypersensitivity to furosemide, torsemide, or related agents All-cause mortality over 30 months (initial trial protocol indicated 12 months of follow up, but patients recruited in the beginning of the trial were continued to be followed until 30 months with 6-month interval telephone contacts. Combination of all-cause mortality and all-cause hospitalizations over 12 months Total number of hospitalizations over 12 months Health-related quality of life (not reported in American Heart Association presentation) Symptoms of depression (not reported in American Heart Association presentation) The TRANSFORM-HF trial was a pragmatic trial. The type of loop diuretic used for maintenance therapy was randomized, but dose and frequency were left to the treating physician. A conversion factor whereby 40 mg of furosemide corresponded to 20 mg of torsemide or 1 mg of bumetanide was used in the trial. However, during the trial execution, the conversion factor torsemide:furosemide was changed from 1:2 to 1:2-4 on the clinicaltrials.gov website. Nevertheless, the furosemide equivalent dose at discharge using an 1:2 conversion factor was very similar in patients treated with torsemide (80 ± 70 mg) vs. furosemide (79 ± 56 mg). The Data and Safety Monitoring Board recommended ending recruitment because of sufficient events (n = 747) after a median follow up of 17.4 months (event rate of 17.0 per 100 patient-years of follow up). Hazard curves for the primary endpoint were almost superimposable with a hazard ratio (95% confidence interval) equal to 1.02 (0.89-1.18) for treatment with torsemide vs. furosemide (P = 0.77). The secondary endpoint of all-cause mortality or any hospital admission was also similar between both groups with an event rate of 107.6 per 100 patient-years in the furosemide group and 99.2 per 100 patient-years in the torsemide group (P = 0.11), corresponding to a hazard ratio (95% confidence interval) of 0.92 (0.83-1.02). There were 987 hospitalizations among 577 patients (40.4%) in the furosemide group vs. 940 hospitalizations among 536 patients (37.5%) in the torsemide group, a difference that was not statistically significant. Results of an on-treatment analysis were in line with the overall results of the main intention-to-treat analysis. Cross-over between treatment arms was observed in 5.4% from randomization to discharge and in 6.7% at 30 days of follow up. At corresponding time intervals, 2.8 and 7% stopped taking their loop diuretic. The results of the TRANSFORM-HF trial do not indicate that maintenance loop diuretic therapy with torsemide compared with furosemide is associated with improved clinical outcome in patients with recent worsening HF. Results with respect to the primary endpoint of all-cause mortality were comparable among both treatment allocation arms. The pragmatic trial design makes the results of this trial very applicable to inform clinical practice as its population reflects a real-world population with worsening HF. Limitations of the trial are: (i) an open-label design, although this is of minor importance with a robust objective endpoint such as all-cause mortality; (ii) a relatively high cross-over rate biasing the results towards the null hypotheses, yet an on-treatment analysis suggests that this probably did not had a major influence on the trial results; (iii) no information on other diuretics prescribed, effectiveness of decongestion, impact on renal function, and electrolyte status, which would all have been very interesting to know; and (iv) there might have been a greater familiarity with furosemide titration among physicians when compared with torsemide as the default diuretic prior to randomization was furosemide in four out of five patients. In addition, the 20% anticipated reduction in all-cause mortality with torsemide was probably overly optimistic for the power calculation of the study. Highly impactful recent randomized clinical trials with sacubitril/valsartan in HFrEF (16% all-cause mortality reduction) or sodium-glucose co-transporter-2 inhibitors in HFpEF (no significant reduction in all-cause mortality) did not achieve this number despite an important impact on cardiovascular mortality and/or HF readmission.4–6 Mechanistic benefits of torsemide utilization not related to the primary endpoint may have been missed due to the pragmatic nature of the trial. Hence, it may remain worthwhile to investigate further the effects of torsemide vs. furosemide on HF-specific outcomes, especially in selected patients with a high risk of diuretic resistance. Until such results become available, achieving complete decongestion in our patients with HF is probably more important than the agent that we prescribe to achieve this goal. Upfront therapy with intravenous acetazolamide inhospital has demonstrated to serve this goal in the recent Acetazolamide in Decompensated Heart Failure with Volume Overload (ADVOR) trial, and is one of the few evidence-based strategies we now have available.7 The ADVOR trial might suggest that sequential nephron blocking should be employed at a lower threshold, especially during the inhospital phase, which might be more effective than altering the dose or type of loop diuretic maintenance therapy out of the hospital. The generation of high-quality evidence on diuretic utilization in HF is however encouraging as the effective treatment of congestion remains a primary focus of any clinician. None declared.