Proteomic differences among patients with heart failure taking furosemide or torsemide

Abstract Background Loop diuretics are commonly used for patients with heart failure (HF) but it remains unknown if one loop diuretic is clinically superior. Hypothesis Biomarkers and proteomics provide insight to how different loop diuretics may differentially affect outcomes. Methods Blood and urine were collected from outpatients with HF who were taking torsemide or furosemide for >30 days. Differences were assessed in cardiac, renal, and inflammatory biomarkers and soluble protein panels using the Olink Cardiovascular III and inflammation panels. Results Of 78 subjects, 55 (71%) were treated with furosemide and 23 (29%) with torsemide, and 25 provided a urine sample (15 treated with furosemide, 10 with torsemide). Patients taking torsemide were older (68 vs 64 years) with a lower mean eGFR (46 vs 54 ml/min/1.73 m2), a higher proportion were women (39% vs 24%) and Black (43% vs 27%). In plasma, levels of hs‐cTnT, NT‐proBNP, and hsCRP were not significantly different between groups. In urine, there were significant differences in urinary albumin, β‐2M, and NGAL, with higher levels in the torsemide‐treated patients. Of 184 proteins testing in Olink panels, in plasma, 156 (85%) were higher in patients taking torsemide but none were significantly different after correcting for false discovery. Conclusions We show differences in urinary biomarkers but few differences in plasma biomarkers among HF patients on different loop diuretics. Olink technology can detect differences in plasma protein levels from multiple biologic domains. These findings raise the importance of defining differences in mechanisms of action of each diuretic in an appropriately powered study.


| INTRODUCTION
The prevalence of heart failure (HF) is increasing in the United States and worldwide. Loop diuretics including furosemide and torsemide remain a foundation of therapy for patients with HF and fluid retention.
While furosemide is much more commonly used, it remains unknown if one loop diuretic is clinically superior for patients with HF. [1][2][3] In these patients with HF, renal and cardiac function are closely linked. Renal function is monitored during treatment as many therapies have the potential to worsen renal function, though the mechanisms and longterm impact of this worsening renal function (WRF) are uncertain. The manner by which torsemide and furosemide may differentially alter the trajectory of renal function and ultimately affect outcomes for patients with HF remains undetermined. 4,5 Soluble protein biomarkers are surrogates of physiological status and risk assessment and are used as clinical cornerstones for the diagnosis and prognosis of patients with HF. However, traditional targeted biomarker measurements of one to several soluble proteins representing a limited number of mechanisms or end end-organ sequela may not be able to identify the mechanistic differences that could provide insight into potential outcome related differences between the diuretics. Until recently, technological limitations and cost restricted the measurement of broad panels of low-abundance proteins in large cohorts and clinical trial studies with the analytical precision to elucidate subtle biologic differences. Targeted discovery proteomics, inclusive of highfidelity assays providing tightly reproducible measures of multiple soluble proteins, provides an opportunity to follow subtle longitudinal changes by measuring low abundance soluble proteins that track with the natural progression of disease or reflect the influence of an intervention. Antibody-oligonucleotide technology by Olink is a novel method of targeted discovery proteomics that is gaining acceptance as an accurate and affordable methodology for high throughput proteomics in large populations. 6 Whether this technology can help inform the mechanism by which torsemide and furosemide affect renal function and outcomes in patients with heart failure remains unknown. In this pilot study, we sought to identify and characterize WRF events during the recovery from acute HF using a panel of urinary biomarkers of tubular injury and examine whether diuretic strategy influences the change in these renal tubular injury biomarkers. Additionally, we sought to compare traditional biomarkers in patients with HF on longterm diuretic therapy and determine the utility of this Olink ELISAbased technology in assessing differences between diuretics to inform future trial design.

| METHODS
Patients with a clinical diagnosis of HF and who were taking either torsemide or furosemide for >30 days, and were willing to provide a blood and urine sample, were enrolled from a single integrated health T-statistics for treatment differences between groups for each protein were calculated. A nominal p-value of <0.05 was considered significant. False discovery rate (FDR) adjusted p-values were calculated accounting for the number of proteins examined and grouping into panels. We controlled for false discovery using the method of Benjamini-Hochberg. 22 FDR p-values of less than 0.05 were considered statistically significant.
The Inova Health System institutional review board approved this study. Prospectively enrolled patients signed an informed consent.
Biobank samples were previously approved for exploratory analysis.

| RESULTS
We analyzed plasma from 78 patients, 52 patients who were prospectively consented and enrolled at 1 of 3 clinic sites and 26 samples from an institution-specific biobank. Of those patients, 25 provided a urine sample (15 treated with furosemide, 10 with torsemide). In this nonrandomized cohort, 55 subjects were treated with furosemide and 23 with torsemide ( Table 1). The mean age was 66 years, with 28% female and 32% black subjects. The torsemide group was older, included a higher proportion of women and Black patients compared to furosemide. Patients taking torsemide also had a lower estimated glomerular filtration rate, and lower systolic blood pressure.
Demographics for the urine cohort were similar with an average age of 66 ± 14 years old, 80% male, and 72% white. The mean estimated glomerular filtration rate was lower in the torsemide group compared to the furosemide group (39.1 ± 11.8 vs 51.4 ± 12.0 ml/min/1.73m 2 ), and mean weight was higher (91.4 vs 88.7 kg). One patient in the furosemide group had marked outliers in several biomarkers and was excluded from the urinary analysis.
In blood, NT-proBNP levels were similar between the torsemide and furosemide group (median 1298 vs 1259 pg/ml, p = 0.84; Table 2). In addition, there was no difference between patients on torsemide and furosemide for biomarkers hsCRP, IL6,  For targeted discovery proteomics for the plasma samples, in the two panels, there were 40 proteins that by t-test, uncorrected for false discovery, were different between patients taking furosemide versus torsemide (Table 3). Despite 40 proteins meeting the threshold for nominal significance, none met significance after correction for false discovery. Box plots for the five proteins with the largest difference are shown in Figure 1. All measured proteins from plasma are shown in a volcano plot in Figure 2. A smaller proportion of the proteins were detected in urine and at lower concentrations than plasma.

| DISCUSSION
In this single center, prospective pilot study of ambulatory patients with HF receiving torsemide or furosemide, we analyzed both blood and urine biomarkers to assess for differences in patients receiving these diuretics and tested the feasibility of utilizing Olink targeted discovery proteomics to detect differences. We demonstrated that there are marked differences in three urinary biomarkers between patients, but only 1 significant differences in many commonly tested blood biomarkers, highlighting the need for methods that can search more broadly and detect more subtle, but potentially meaningful differences between patients on these drugs. In exploring the Olink methodology of targeted discovery proteomics, we showed that this technology can potentially detect subtle differences between groups in plasma but with a low yield of protein detection in urine.
While there is substantial supportive evidence for many guideline-directed therapies in HF, there are limited data regarding one of the most commonly used classes of drugs in HF, loop diuretics. 23 Despite suggestive preclinical mechanistic and clinical data, robust randomized controlled trial data are only now being generated. [24][25][26][27][28][29] Loop diuretics have been shown to activate the renin-angiotensinaldosterone system (RAAS), leading to deleterious effects and progression of HF. 30 However, while furosemide increases circulating aldosterone levels, torsemide has anti-aldosteronergic and antikaliuretic effects, with reductions in aldosterone secretion and receptor binding, leading to improvements in ventricular wall stress, decreased sympathetic nervous system activity, direct vasodilation, and multifactorial reductions in detrimental myocardial fibrosis. [31][32][33][34][35][36][37][38][39][40][41] Additionally, torsemide has better bioavailability and potency than furosemide which may contribute to superior efficacy, and outcomes including improved functional status, quality of life, and reduced hospitalizations. [42][43][44] The TORIC study, the largest study to date comparing torsemide and furosemide, was an open-label post-marketing surveillance study; although it showed more functional improvement and less mortality in patients receiving torsemide, these results must be taken in the context of the non-randomized study design. 45 Despite suggestions from open-label trials that patients treated with torsemide have better outcomes, results from other studies demonstrated no benefit. 27,28 Two recent metaanalyses showed that torsemide was associated with improved F I G U R E 1 Select box and whisker plots for differences in proteins from patients on furosemide versus torsemide functional status compared to furosemide, but clinical outcomes were similar between the diuretics. 25,29 Renal function likely plays a role in clinical outcomes for HF patients on diuretic therapy. In the acute HF setting, WRF has been associated with adverse outcomes independent of co-morbidities and LVEF. 2  of immunoassays that could be tested on blood samples from large cohort studies. Discovery proteomics by mass spectroscopy, which has been available for some time, can be used to discover thousands of proteins. However, mass spectroscopy has limited sensitivity to detect many of the cardiac and renal related mechanistic proteins of interest that exist in low abundance in plasma (i.e., concentrations of pg/ml) Moreover, a notable limitation for mass spectroscopy when using blood is that higher abundance proteins (i.e., albumin) typically need to be removed to detect lower abundance proteins, but many low-abundance proteins bind to the higher abundance proteins and may be co-removed in the process. 38 Recently, two novel technologies have been validated in cardiovascular disease to measure hundreds to thousands of proteins. [39][40][41][42][43] The single aptamer based detection approach has the potential to provide a broad coverage of targeted discovery proteins, but recent concerns have been raised with respect to the specificity associated with this technology. 44 This current study evaluated the feasibility of the other novel technology, antibody-oligonucleotide technology by Olink. We have shown the Olink technology is capable of detecting differences in plasma in protein concentrations from multiple biologic domains relevant to cardiac pathophysiology.
Our study had several limitations. Given the cross-sectional design, modest sample size, and likelihood of confounding by indication, it is not possible to determine that torsemide results in higher levels of blood and urine soluble inflammatory and fibrosis proteins.
The urine biomarker analysis was further limited by lack of available urinary concentration measurements. However, these data raise the importance of defining differences in biologic sequela based on mechanisms of action of each diuretic in an appropriately powered ancillary study within a randomized controlled trial.
In conclusion, this analysis shows that there are limited differences in plasma biomarkers but several significant difference in some urinary biomarkers among HF patients on chronic furosemide therapy versus torsemide therapy. Further, the Olink technology is capable of detecting differences in protein concentrations from multiple biologic domains relevant to cardiac pathophysiology in patients on torsemide compared with furosemide. Further study is needed to elucidate the underlying mechanisms for the differences between these groups.

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of the article at the publisher's website.