Mechanistic Differences between Torsemide and Furosemide

Visual Abstract

Abstract

Key Points

• Oral torsemide was not superior to furosemide in measures of renal tubular delivery or duration of action.

• A dose equivalence of approximately 40 mg oral furosemide:10 mg oral torsemide resulted in similar natriuresis.

• The two-fold higher doses of torsemide did not improve fluid status due to the kidney’s compensation.

Background

Torsemide is proposed to have clinically important pharmacokinetic and pharmacodynamic advantages over furosemide. However, clinical outcomes did not differ in the Torsemide Comparison with Furosemide for Management of Heart Failure (TRANSFORM) randomized trial.

Methods

We conducted a multicenter mechanistic substudy of patients with heart failure randomized to oral furosemide or torsemide (TRANSFORM-Mechanism trial). At baseline and 30 days, participants underwent detailed assessments of pharmacokinetic and pharmacodynamic parameters.

Results

The TRANSFORM-Mechanism trial enrolled 88 participants. Kidney bioavailability, or the proportion of dose delivered to the tubular site of action, was significantly less with torsemide compared with furosemide (median, 17.1% [interquartile range, 12.3%–23.5%] versus 24.8% [16.6%–34.1%], P < 0.001). Furosemide had a longer duration of kidney drug delivery and natriuresis (P ≤ 0.004 for both). Prescribed doses of furosemide and torsemide in the TRANSFORM-Mechanism trial were similar to the TRANSFORM trial, with clinicians on average using a 2:1 dose equivalence conversion between drugs. However, these doses resulted in a substantially greater natriuresis with torsemide (P < 0.001). A dose equivalence of approximately 4:1 resulted in similar natriuresis. Higher diuretic doses in the torsemide group resulted in mild perturbations in kidney function and significant increases in renin, aldosterone, and norepinephrine (P < 0.05 for all). Plasma volume (P = 0.52) and body weight (P = 0.89) did not improve with torsemide versus furosemide.

Conclusions

We observed no meaningful pharmacokinetic/pharmacodynamic advantages for torsemide versus furosemide. The greater natriuresis from higher diuretic doses in the torsemide group was offset by greater neurohormonal activation and kidney dysfunction.

Clinical Trial registry name and registration number:

TRANSFORM-HF: ToRsemide compArisoN With furoSemide FORManagement of Heart Failure (TRANSFORM-HF), NCT03296813; Torsemide Comparison With Furosemide for Management of Patients With Stable Heart Failure (TFO), NCT05093621.

Introduction

Oral loop diuretics are foundational medical therapies for heart failure, but substantial pharmacokinetic and pharmacodynamic differences are believed to exist between individual members of the class.^1–3 Specifically, torsemide has been proposed to have important pharmacokinetic/pharmacodynamic advantages over furosemide, motivating the Torsemide Comparison with Furosemide for Management of Heart Failure (TRANSFORM-HF) trial.³ However, no significant differences were observed in all-cause death or rehospitalizations between patients randomized to torsemide or furosemide in the TRANSFORM-HF trial.⁴

Given the pragmatic design of the TRANSFORM-HF trial, the mechanisms underlying the lack of benefit are unknown. Importantly, it has yet to be demonstrated that the widely popularized pharmacokinetic/pharmacodynamic advantages of torsemide provide any true clinical value. For example, while torsemide's superior oral bioavailability results in more reliable delivery of drug to the blood, only the loop diuretic delivered into the luminal aspect of the kidney tubules is relevant to its diuretic effect.^5–7 Unlike furosemide, most of absorbed torsemide is metabolized by highly variable hepatic pathways, thus kidney bioavailability may not be superior to furosemide.^8,9 In addition, although the torsemide molecule has a longer terminal elimination t_1/2, this may be negated by oral furosemide's absorption-limited kinetics, where absorption is slower than elimination, acting similar to a sustained-release drug.⁵ Moreover, prescribing physicians in the TRANSFORM-HF trial predominantly used a 2:1 dosing equivalency between torsemide and furosemide,⁴ but the literature supports a 4:1 dosing equivalence in patients with heart failure.^2,5,10 Thus, significant differences in loop diuretic exposure between randomized groups may have been present. The purpose of this investigation was to thoroughly evaluate the pharmacokinetic/pharmacodynamic differences between torsemide and furosemide to allow insights into why a lack of benefit was observed with torsemide in the TRANSFORM-HF trial.

Methods

We conducted a multicenter, prospective, observational, mechanistic study evaluating patients with heart failure who had been randomized 1:1 to oral furosemide versus oral torsemide. Randomization was independent of this mechanistic observational study and occurred during participation in two independent parent studies. The mechanistic study used a single protocol regardless of the source of randomization. Patients were enrolled (1) through the TRANSFORM-HF trial (NCT03296813), which randomized hospitalized patients upon transition from intravenous (IV) to oral diuretics, and (2) through the Torsemide Comparison with Furosemide for Management of Patients with Stable Heart Failure (TRANSFORM-Outpatient) (NCT05093621) parent study, which enrolled stable outpatients. Patients enrolling in either TRANSFORM parent study at the Yale University and University of Utah sites were approached for enrollment in the TRANSFORM-Mechanism trial. The dose and frequency of the loop diuretic therapy during the study were determined by the treating clinician for both parent studies, with a guidance that between 2 and 4 mg of furosemide was equivalent to 1 mg of torsemide, but clinicians largely used a 2:1 dosing conversion in both studies.⁴

The TRANSFORM-HF trial's methods and results have been previously described.³ In brief, the TRANSFORM-HF trial was a prospective, randomized, pragmatic, comparative-effectiveness study comparing the effect of torsemide versus furosemide on long-term mortality, hospitalization, and patient-reported outcomes among patients with heart failure. The pragmatic design enrolled patients with worsening chronic or new heart failure and need for chronic daily outpatient oral loop diuretic therapy. The inclusion and exclusion criteria for patients enrolled through the TRANSFORM-Outpatient trial were largely the same as those for the TRANSFORM-HF trial apart from the removal of heart failure hospitalization–related inclusion criteria and addition of a stable oral loop diuretic dose for the past 30 days as an inclusion criterion (Supplemental Table 1). The TRANSFORM-Outpatient parent study was added to improve the slow rate of enrollment during the coronavirus disease 2019 pandemic. This change also served to improve generalizability of the findings to the overall TRANSFORM-HF exposure, which was primarily remote from the index hospitalization.⁴ These studies were approved by each site's institutional review board, and all patients provided written informed consent.

Study Visit

Study visit procedures were identical between patients enrolled through inpatient and outpatient pathways. We conducted an approximately 10-hour study visit, including an 8-hour timed urine collection protocol at the time of first study diuretic dose administration (Supplemental Figure 1). In prior pharmacokinetic/pharmacodynamic studies, most of the urinary drug excretion and diuretic action are complete for both diuretics by 8 hours.^5,11,12 After an overnight fast, patients were instructed to take their chronic medications as usual except the oral study diuretic. After 1 hour of quiet recumbency, the angle of the bed was noted for consistency between study visits and a blood sample was collected. Next, blood volume was measured using indicator dilution with I-131 albumin (Daxor Inc., New York, NY). The oral study diuretic was then administered at hour 0. Blood and spot urine samples were collected at hours 1, 2, 4, 6, and 8, with cumulative urine collections occurring between each time point totaling an 8-hour urine collection. Timed urine collections were ended with a forced void and confirmed by a bladder ultrasound. In patients with logistical barriers to participating in a 10-hour study visit (n=29/88, 33%), the study visit was shortened to 6 hours, terminating urine collection after the hour 4 urine and blood collection, but was otherwise identical. After the study visit, patients performed a separate, unsupervised cumulative urine collection at home for the remainder of the day to complete 24-hour urine collection to quantify 24-hour natriuresis after diuretic administration. We repeated identical study procedures on day 30.

Study Measures

We measured the pharmacokinetic profile of torsemide and furosemide by evaluating the excretion of unchanged drug in the urine over the study visits, as the metabolism of both drugs is complete prior to the loop of Henle.^13,14 Kidney bioavailability was defined as the percentage of the oral dose that was recovered unchanged in urine over the study visit and, therefore, reflects the percentage of active drug reaching the nephron tubular lumen site of action. Furosemide and torsemide in urine were measured using liquid chromatography–mass spectrometry per our established protocol¹⁵ (Supplemental Analytical Methods). Twenty-five individual blinded patient samples were diluted and run in duplicate; the interassay coefficient of variation was 6.1%±4.2% for furosemide and 4.1%±2.1% for torsemide. Norepinephrine and aldosterone were measured with the commercially available ELISA kit from ALPCO according to the manufacturer's instructions (ALPCO, Salem, NH). Total renin was measured with ELISA kits from R&D Systems (Minneapolis, MN). We measured plasma amino-terminal propeptide of type III procollagen using the commercial type III procollagen ELISA kit from Cisbio (Perkin Elmer) as per manufacturer's instructions (Supplemental Analytical Methods).

We measured the pharmacodynamic profiles using cumulative natriuresis and fractional excretion of sodium (FENa) across all study visit time points (Supplemental Analytical Methods). eGFR was calculated with cystatin-based and creatinine-based Chronic Kidney Disease Epidemiology formulas using the race-free equation. Diuretic efficiency was used to compare diuretic responsiveness in the setting of differences in prescribed doses. Diuretic efficiency was defined as natriuretic response normalized for the log based 2 of the diuretic dose administered during the study visit, as previously described and validated.¹⁵

Statistical Analyses

Data that were approximately normally distributed are presented as mean±SD, and data with skewed distribution are shown as median (quartile 1– 3). Categorical values are presented as frequency and percentage. Continuous data with normal or skewed distribution were compared with Student t test or Mann–Whitney U test, respectively. Categorical data were compared with chi-square test. Unadjusted logistic regression was used to analyze the odds of an increase in renin, aldosterone, and norepinephrine. For comparisons of loop diuretic kidney bioavailability, FENa, total sodium excretion, changes in biomarkers (e.g., norepinephrine, renin, aldosterone), weight, plasma volume, and total blood volume, we accounted for the absence of independence of observations with linear mixed models. Patients completing the 6-hour study visit are included in all analyses and up to hour 4 in time-based analyses. A detailed description of the statistical tests used in the article, including how inability to urinate at a time point was analyzed, is found in Supplemental Statistical Analysis Methods. We prespecified analyses of patient characteristics and outcomes between patients enrolled through the TRANSFORM-HF versus TRANSFORM-Outpatient trial. Statistical significance was defined as two-tailed P < 0.05. Statistical analyses were performed using Stata SE version 17 (Statacorp, College Station, TX).

Results

A total of 88 patients were enrolled between July 2019 and March 2022 before the TRANSFORM-HF trial was stopped. Baseline characteristics are presented in Table 1 and were similar to the parent TRANSFORM-HF trial (Supplemental Table 2) across most characteristics and between patients enrolled through the TRANSFORM-HF (n=41, 47%) and TRANSFORM-Outpatient (n=47, 53%) trials (Supplemental Table 3). The number of patients receiving once-daily diuretic dosing (76% torsemide and 73% furosemide, P = 0.63) were similar between treatment arms. Supporting the a priori pooling of the two enrollment sources, there was no significant heterogeneity in pharmacokinetic/pharmacodynamic parameters between the patients randomized in the TRANSFORM-HF and TRANSFORM-Outpatient enrollment sources with respect to loop diuretic dose (P interaction = 0.70), kidney bioavailability (P interaction = 0.78), total visit natriuresis (P interaction = 0.71), FENa (P interaction = 0.73), plasma volume (P interaction = 0.28), and total blood volume (P interaction = 0.90). There was no significant heterogeneity for loop diuretic treatment arm (P = 0.94) or dose (P = 0.74) between the study visit lengths, and the observed differences between treatment arms were consistent between study visit lengths for kidney bioavailability (P interaction=0.26), total visit natriuresis (P interaction = 0.61), and FENa (P interaction = 0.68).

Table 1.

Baseline characteristics of participants in a pharmacokinetic and pharmacodynamic study of torsemide versus furosemide

Characteristics	Total Cohort N=88	Torsemide n=42 (48%)	Furosemide n=46 (52%)
Age, yr	61±13	61±11	60±14
Male	63 (72)	31 (74)	32 (70)
Race
Black	32 (36)	14 (33)	18 (39)
Other	4 (5)	3 (7)	1 (2)
White	52 (59)	25 (60)	27 (59)
BMI, kg/m2	35±13	37±17	34±7
LVEF, %	39±16	42±17	37±15
LVEF >50%	27 (31)	15 (37)	12 (26)
Ischemic cardiomyopathy	24 (27)	13 (31)	11 (24)
Heart failure hospitalization in the past year	22 (25)	10 (24)	12 (26)
In-hospital enrollment	41 (47)	18 (43)	23 (50)
Comorbid conditions
Hypertension	66 (75)	34 (81)	32 (70)
Diabetes mellitus	39 (44)	20 (48)	19 (41)
History atrial fibrillation/flutter	32 (36)	14 (33)	18 (39)
Vital signs
Systolic BP, mm Hg	125±20	128±19	122±19
Diastolic BP, mm Hg	77±11	79±10.8	75±10
Heart rate, bpm	73±13	74±12	73±14
Medications
ACEi or ARB or ARNI	59 (67)	26 (62)	33 (72)
Beta blocker	76 (86)	36 (86)	40 (87)
Mineralocorticoid receptor antagonist	36 (41)	14 (33)	22 (48)
Daily dose, mg	25 (25–25)	25 (25–25)	25 (22–25)
SGLT2i	26 (30)	15 (36)	11 (24)
Hydralazine and nitrates	6 (7)	2 (5)	4 (9)
Prerandomization chronic oral loop diuretic	72 (82)	32 (76)	40 (87)
Thiazide	1 (1)	0 (0)	1 (2)
Laboratory values
Serum sodium, mEq/L	139±4	139±3	139±4
Serum chloride, mEqL	101±3	101±3	101±3
Serum bicarbonate, mEq/L	22±3	23±4	22±3
Serum creatinine, mg/dl	1.2±0.4	1.2±0.3	1.1±0.4
BUN, mg/dl	23±10	23±8	22±12
eGFR, ml/min per 1.73 m2	64±25	62±26	67±24
eGFR <60 ml/min per 1.73 m2	44 (50)	23 (54)	21 (46)
eGFR <30 ml/min per 1.73 m2	5 (6)	3 (7)	2 (4)
NT-pro-B-type natriuretic peptide, pg/ml	1694 (440–4207)	1544 (660–3599)	1825 (271–4636)

Values presented as number (%), mean±SD, or median (interquartile range) as appropriate. ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; ARNI, angiotensin receptor–neprilysin inhibitor; BMI, body mass index; bpm, beats per minute; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal pro-B-type natriuretic peptide; SGLT2i, sodium-glucose cotransporter-2 inhibitor.

Pharmacokinetics

Kidney bioavailability, or the percentage of the oral dose recovered in urine during the timed urine collection and, therefore, the amount of drug reaching the tubular site of action, was significantly higher for furosemide than torsemide (median, 24.8% [interquartile range, 16.6%–34.1%] versus 17.1% [12.3%–23.5%], P < 0.001) (Figure 1A). In line with furosemide's absorption-limited pharmacokinetics, the temporal pattern of loop diuretic delivery to urine was consistent with a slower onset and prolonged delivery to urine with furosemide versus torsemide (P = 0.003, Figure 1B). Torsemide had a higher percentage of the dose excreted in urine in the first 2 hours (P < 0.001), whereas furosemide had a higher percentage of the dose excreted after 2 hours (P = 0.003).

Figure 1.

Pharmacokinetics of oral furosemide and torsemide. (A) Furosemide displayed greater kidney bioavailability, defined as the percentage of the oral dose excreted unchanged in urine during the timed urine collection (i.e., diuretic delivered to the site of action), compared with torsemide. (B) Kinetics of diuretics reaching the urinary site of action displayed as the percentage of total kidney delivery over time. Furosemide had a significantly slower initial onset of delivery to urine, followed by a longer duration of delivery. The number of urine samples analyzed at each time point, and when a urine sample was not available, the median volume of urine in the bladder (quantitated by formal bladder ultrasound) was HR1: n=134, 28 ml (1–90 ml); H2: n=123, 26 ml (0–181 ml); HR4: n=132, 41 ml (0–98 ml); HR6: n=97, 1 ml (0–36 ml); and HR8: n=91, 37 ml (3–98 ml). Results were consistent across various sensitivity analyses (Figure 1B and Supplemental Statistical Analysis Methods). Data are presented as median (IQR) and mean±SEM of the mean in (A) and (B), respectively. (C) Population variability in kidney bioavailability between patients expressed as interleaved histograms. (D) Variability in kidney bioavailability within individual patients from baseline to the 30-day visit. (C and D) Standardized data with mean (0) and SDs from the mean; P values represent comparisons of distributions between treatment groups. IQR, interquartile range.

Variability in Kidney Bioavailability

Substantial variability in kidney bioavailability both between patient and within patient was evident with both drugs (Figure 1, C and D). Overall, the variability tended to be less with torsemide than with furosemide. However, this variability in bioavailability was not a major driver of diuretic response because overall variability in kidney bioavailability with both drugs accounted for only 31% (95% confidence interval [CI], 4% to 51%) of the observed variability in natriuresis at the population level and 11% (95% CI, 1% to 32%) for the difference in natriuresis within the same patient between study visits. The variability in natriuresis at the population level (difference in each patient's diuretic efficiency versus the population median) and within patient (difference in diuretic efficiency between the baseline and 30-day visit) was not different for torsemide versus furosemide (P ≥ 0.35 for both).

Dosing Equivalency

Consistent with the TRANSFORM-HF study, physicians predominantly used a 2:1 conversion for furosemide (mean 50±30 mg, median 40 [20–80] mg) to torsemide (mean 26±19 mg, median 20 [20–40] mg) (Figure 2A). Because participants were randomized, diuretic responsiveness should be similar between groups at baseline. Thus, equivalent natriuresis would have been observed if 2:1 was the correct dose equivalent conversion. However, natriuresis was substantially greater in the torsemide group (FENa 4.1%±3.5% versus 2.5%±2.4%, P < 0.001, Figure 2B). Figure 2C demonstrates that as the oral dose equivalent ratio is varied from 2:1 to 5:1, equivalent natriuresis occurs when the conversion ratio approaches 4:1.

Figure 2.

Oral diuretic dose equivalence. (A) Quantity of loop diuretic administered using a 2:1 equivalence ratio for furosemide to torsemide was similar. (B) However, the observed natriuresis was substantially greater with torsemide. (C) The oral dose equivalence ratio of furosemide to torsemide is varied from 2:1 to 5:1; diuretic response is equivalent with approximately a 4:1 ratio. *FENa diuretic efficiency represents the increase in FENa per 2× increase in diuretic dose.¹⁶ Data are presented as median (IQR) (A and B) and mean±SEM of the mean (C). FENa, fractional excretion of sodium.

Natriuresis

Because torsemide was prescribed at a dose approximately twice as potent as furosemide, the observed natriuretic response was consistently greater for torsemide (Figures 2B and 3A). Torsemide produced significantly greater mean cumulative natriuresis than furosemide during the study visit (Figure 3A). Similar to their pharmacokinetic profiles, torsemide produced a greater percentage of the associated total natriuresis during the first 2 hours (P = 0.015), whereas after hour 2, furosemide had a larger percentage of total natriuresis (P = 0.004) (Figure 3B). Torsemide resulted in an additional 44 mmol (95% CI, 25 to 62 mmol; P < 0.001) of absolute natriuresis compared with furosemide from hour 0 to 4 (58% more than furosemide), but furosemide had nominally greater natriuresis (6 mmol; 95% CI, −8 to 19 mmol) from hour 4 to 8 (13% more than torsemide), which was not significantly different (P = 0.56).

Figure 3.

Pharmacodynamics of oral furosemide and torsemide. (A) Torsemide produced a greater cumulative sodium excretion (mmol) over the entire study visit. (B) Natriuresis over time with each drug is shown. Torsemide produced a greater proportion of its total sodium excretion during the first 2 hours compared with furosemide. After hour 2, furosemide produced a greater proportion of its natriuresis compared with torsemide. Data are presented as median (IQR) (A) and mean±SEM of the mean (B).

Change in Diuretic Response from Baseline to 30 Days

Changes in diuretic dose were not statistically significant from baseline to 30 days between torsemide and furosemide (P = 0.27). With the first randomized exposure to a diuretic, torsemide resulted in a significantly greater acute natriuresis over the study visit (furosemide 97±64, torsemide 151±87, P = 0.002). However, at the 30-day study visit, torsemide was no longer associated with significantly greater natriuresis (furosemide 111±79, torsemide 140±77, P = 0.10). Although unsupervised and thus likely with errors in collection, 24-hour cumulative sodium output tended to be less with furosemide (181±104 mmol) versus torsemide (236±154) at the randomization visit (P = 0.08). At the 30-day visit, there was no difference in 24-hour cumulative sodium output between groups (furosemide 201±82 versus torsemide 205±132, P = 0.88). Time-by-treatment interactions were not significant for the above analyses (P > 0.14 for all).

Neurohormonal Activation and Kidney Function

Patients randomized to torsemide had significant increases in norepinephrine (P = 0.039), renin (P < 0.001), and aldosterone (P = 0.002) from baseline to 30 days compared with those to furosemide (Figure 4). The odds for an increase in norepinephrine (odds ratio [OR], 2.9; 95% CI, 1.1 to 7.4), aldosterone (OR, 1.7; 95% CI, 0.7 to 4.2), or renin (OR, 3.3; 95% CI, 1.3 to 8.3) or an increase in all three (OR, 3.4; 95% CI, 1.1 to 10.9) was higher with torsemide. These associations were essentially unchanged after multivariable adjustment for concomitant neurohormonal antagonist medications. Changes in blood chemistries over the study period are presented in Table 2. Torsemide caused a small worsening in BUN, serum creatinine, BUN–creatinine ratio, uric acid, and cystatin C compared with furosemide. Increase in serum sodium was similar between groups, but serum chloride increased significantly more with furosemide, and bicarbonate tended to worsen with torsemide.

Figure 4.

Change in plasma norepinephrine, total renin, and aldosterone from baseline to 30 days. Data are presented as median (IQR).

Table 2.

Change in blood chemistries over 30 days

Laboratory Value	Torsemide	Furosemide	P Value
Sodium, mEq/L	0.9±3.7	0.8±3.4	0.87
Potassium, mEq/L	−0.2±0.6	−0.1±0.5	0.40
Chloride, mEq/L	0.2±4.6	1.0±3.4	0.003
Bicarbonate, mEq/L	0.5±3.5	−0.1±3.4	0.06
BUN, mg/dl	2.4 (−0.7 to 7.3)	−0.6 (−3.5 to 2.5)	<0.001
Serum creatinine, mg/dl	0.02 (−0.08 to 0.13)	−0.01 (−0.07 to 0.09)	0.03
BUN/serum creatinine ratio	2.6±5.8	−0.7±4.9	<0.001
Cystatin-C, mg/L	0.04±0.3	0.01±0.3	0.13
Uric acid, mg/dl	0.4±2.0	−0.2±1.2	<0.001
Magnesium, mg/dl	−0.05±0.46	−0.08±0.34	0.002
Phosphate, mg/dl	−0.1±0.7	0.0±0.7	0.08
Calcium, mg/dl	0.0 (−0.3 to 0.4)	0.0 (−0.2 to 0.2)	0.75

Values are presented as mean±SD or median (interquartile range).

Changes in Volume Status

The change in volume status between baseline and 30 days primarily reflects the area under the curve of sodium balance over that time. The change between baseline and 30 days in total blood volume (P = 0.38) and plasma volume (P = 0.52) was not significantly different with torsemide compared with furosemide (Figure 5). Weight change was also not different between drugs (Figure 5).

Figure 5.

Change in blood volume, plasma volume, and body weight from baseline to 30 days. Data are presented as median (IQR).

Potential Antialdosterone Effects

As reported above, plasma aldosterone increased significantly in the torsemide group, suggesting that torsemide's potential aldosterone synthase–inhibiting effects were insufficient to outweigh the higher diuretic dose in the torsemide group. We also were unable to detect the proposed mineralocorticoid receptor antagonistic effect of torsemide (Supplemental Figure 2).

Discussion

The primary findings of the TRANSFORM-Mechanism study evaluating oral torsemide and oral furosemide in patients with heart failure are as follows: (1) Kidney bioavailability was greater with oral furosemide versus torsemide and similarly variable between drugs. (2) As predicted by furosemide's absorption-limited kinetics, a longer duration of both kidney drug delivery and diuretic action was observed with furosemide compared with torsemide. (3) A 2:1 dose conversion conclusively does not result in equivalent natriuresis between oral furosemide and oral torsemide. Rather, this study supports the oral dose equivalence of 4:1 previously reported in the literature.^2,5,10 (4) Physicians used a 2:1 dose equivalence, causing a greater natriuresis with torsemide. (5) Likely related to these differences in dose and natriuresis, randomization to torsemide caused neurohormonal activation, mildly worsening kidney function, and uric acid increases. (6) Kidney compensation to the higher dose of torsemide ultimately resulted in an absence of differences in blood volume or body weight at 30 days compared with furosemide. (7) If torsemide possesses any antialdosterone effects, the magnitude was too small to offset the differences in dosing between randomized groups. Collectively, these findings indicate that torsemide lacks clinically significant pharmacokinetic/pharmacodynamic superiority over furosemide when used at equivalent doses, but relatively modest differences in dosing can trigger neurohormonal activation and kidney compensation.

Torsemide has two primary proposed pharmacokinetic advantages over furosemide: (1) superior and more consistent oral bioavailability and (2) a significantly longer elimination t_1/2.^3,6,7 These two differences are well-established facts about the torsemide versus furosemide molecules. However, these are not true advantages when considering the relevant biology related to diuresis. Torsemide's superior oral bioavailability only considers delivery of drug to blood, which alone is an irrelevant parameter for diuresis because it is drug delivery to the luminal aspect of the kidney tubular epithelium that elicits natriuresis.⁶ Torsemide had inferior absolute kidney bioavailability, likely because of torsemide's metabolism by variable hepatic pathways, rather than kidney excretion.^5,8,17 Importantly, the consistency of kidney bioavailability was not meaningfully better with torsemide and did not translate into more consistent diuretic effects. Overall, these findings indicate that the amount or consistency of drug delivered to the site of action is not meaningfully different between the two diuretics.

In addition to lacking superior absolute kidney drug delivery, torsemide's duration of kidney delivery was inferior to furosemide. Unlike torsemide, furosemide has absorption-limited kinetics where the rate of gastrointestinal absorption is slower than the rate of elimination, minimizing the relevance of furosemide's shorter t_1/2 during oral administration.⁵ Conversely, furosemide's slower absorption may be an advantage. Compared with the commonly prescribed immediate-release torsemide, a new extended-release torsemide preparation with slower absorption produced greater natriuresis and reduced worsening kidney function, despite an overall reduction in gastrointestinal bioavailability.¹¹ Thus, the slow onset and prolonged duration of action of oral furosemide may result in a better cardiorenal and natriuretic profile in some situations.

This study provides convincing evidence that a 2:1 oral dose conversion between furosemide and torsemide does not provide equivalent natriuresis and supports prior literature that a 4:1 dose equivalence of oral furosemide to oral torsemide is appropriate in heart failure.^2,5,10 Two randomized, crossover trials of patients with heart failure found oral torsemide 10 mg had equivalent natriuresis to oral furosemide 40 mg.^5,10 In addition, the totality of the literature strongly supports a 4:1 conversion. It is well established that a 2:1 dosing equivalence produces similar natriuresis between IV furosemide and IV torsemide.^2,6,17–20 On average, approximately 50% of oral furosemide is absorbed, whereas 80%–100% of oral torsemide is absorbed.^2,7,21–24 In agreement with oral bioavailability data, a 1:1 IV to oral dosing equivalence has been established for torsemide, whereas a 1:2 IV to oral dosing equivalence is established for furosemide.^{2,5,23,25–31} Coalescing these facts, the literature supports 20 mg IV torsemide=20 mg oral torsemide=40 mg IV furosemide=80 mg oral furosemide, and thus, a 4:1 oral dose conversion for furosemide to torsemide in heart failure should be used.

The TRANSFORM studies were open label, and patient or clinical bias could have affected the results. Diuretic dosing was at the discretion of the treating physician. Patients recently hospitalized and stable outpatients were included, which could affect the pharmacodynamic results, although no differences were observed between these groups. The variability in dietary sodium intake, ability of some patients to only participate in abbreviated study visits, and unsupervised 24-hour urine collections are limitations. Although patients were randomized to treatment groups and baseline characteristics were not different, unmeasured imbalances in variables affecting diuretic resistance between groups could exist given the modest sample size.

Oral torsemide did not display clinically meaningful superiority in pharmacokinetics or pharmacodynamics to oral furosemide. A dose equivalence of 4:1 oral furosemide to oral torsemide produces equivalent natriuresis. However, physicians used a 2:1 dosing equivalency, resulting in an approximate two-fold higher dose in the torsemide arm. This resulted in neurohormonal activation and mild worsening of worsened kidney function with torsemide. Overall, this kidney/neurohormonal compensation resulted in the lack of improvement in volume status despite the greater diuretic dose. These findings highlight the complex risks and benefits associated with chronic diuretic dosing in patients with heart failure.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/E835.

Funding

J.M. Testani: National Heart, Lung, and Blood Institute (NHLBI) Division of Intramural Research (National Institutes of Health [NIH] NIH R01HL139629, R01HL148354, R01DK130997, and R01DK130870) and NHLBI Division of Intramural Research (U01-HL125478 and U01-HL125511). F.P. Wilson: NHLBI Division of Intramural Research (R01DK113191 and P30DK079210).

Author Contributions

Conceptualization: Jeffery M. Testani.

Formal analysis: Juan B. Ivey-Miranda, Julieta Moreno-Villagomez, Daniela Ramos-Mastache, Jeffery M. Testani.

Investigation: Sean P. Collins, David H. Ellison, James C. Fang, Robert J. Mentz, W.H. Wilson Tang, Jeffery M. Testani, Jeffrey M. Turner, Eric J. Velazquez, Christopher S. Wilcox, F. Perry Wilson.

Methodology: Natasha Balkcom, Daniel Neville, Veena S. Rao.

Writing–original draft: Zachary L. Cox, Juan B. Ivey-Miranda, Veena S. Rao, Jeffery M. Testani.

Writing–review & editing: Lavanya Bellumkonda, Sean P. Collins, Zachary L. Cox, David H. Ellison, James C. Fang, Juan B. Ivey-Miranda, Christopher Maulion, Robert J. Mentz, Veena S. Rao, W. H. Wilson Tang, Jeffery M. Testani, Jeffrey M. Turner, Eric J. Velazquez, Christopher S. Wilcox, F. Perry Wilson.

Data Sharing Statement

All data are included in the manuscript and/or supporting information.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/E834.

Supplemental Analytical Methods

Supplemental Statistical Analysis Methods

Supplemental Table 1. Comparison of inclusion and exclusion criteria.

Supplemental Table 2. Baseline characteristics of TRANSFORM-HF and TRANSFORM-mechanism.

Supplemental Table 3. Baseline characteristics and outcomes by inpatient versus outpatient enrollment pathways.

Supplemental Figure 1. Study visit diagram.

Supplemental Figure 2. Measures of potential “anti-aldosterone” effects.