Unsolved challenges in diuretic therapy for acute heart failure: a focus on diuretic response

Abstract

Loop diuretics represent the mainstay of management of patients hospitalized for heart failure (HF). Diuretic resistance is commonly encountered in clinical practice, but limited evidence-based approaches are available to address it. Recent clinical investigations have proposed common definitions of diuretic response: a change in body weight, net fluid loss or total urinary output to 40 mg of furosemide dose equivalents. Poor diuretic response is characterized by features of advanced HF and atherosclerosis and is independently associated with poor in-hospital and post-discharge outcomes. A number of adjunctive or combination decongestion therapies are available to overcome diuretic resistance, but high-quality prospective data supporting these approaches are lacking. Once a definition has been standardized and accepted, diuretic response may represent an important inclusion criteria and end point in upcoming clinical trials in hospitalized HF to help define an optimal, tailored approach to this challenging clinical entity.

Keywords:

• clinical trials
• heart failure
• loop diuretics
• outcomes
• rehospitalizations

Hospitalization for heart failure (HF) accounts for a tremendous global economic and clinical burden [1]. Inpatient pharmacologic management is often limited to loop diuretics alone with low rates of introduction of new therapies during hospitalization [2]. Despite the widespread utilization of loop diuretics worldwide, their use has largely been empiric with high variability across clinical settings. Indeed, loop diuretics have only recently been systematically and rigorously studied.

‘Diuretic resistance’, broadly defined as the “failure to decongest despite adequate and escalating doses of diuretics” [3], has emerged as a frequently encountered clinical scenario. More specifically, diuretic resistance in the context of HF refers to the failure to effectively reduce extravascular volume and affect a negative total sodium balance. The pathophysiology of this phenomenon is likely diverse, reflecting cardiorenal interactions and renal adaptations to ongoing diuretic therapy [4]. The ability of diuretics to promote ongoing negative sodium and fluid balance may be deterred by impaired absorption, bioavailability and efficient diuretic delivery; upregulation of counterregulatory neurohormonal pathways (namely, renin–angiotensin–aldosterone and sympathetic nervous system) and renal-specific adaptive mechanisms that manifest in the diuretic braking and distal tubular hypertrophy and remodeling [4].

The identification of diuretic resistance in clinical practice has been primarily qualitative and clinician-dependent. Recent clinical studies have attempted to standardize the definition of diuretic response or efficiency, based on readily available metrics. Reduced diuretic response has been shown to be an independent marker of poor in-hospital and post-discharge prognosis in patients hospitalized for HF. Clear identification of the subset of patients who exhibit poor diuretic response may facilitate tailored therapeutic strategies with adjunctive or combination pharmacotherapies. In this brief piece, we review recent clinical data adding precision to the definition of diuretic response and discuss its potential use in upcoming acute HF clinical trials.

Diuretic response: approaching a definition

The pharmacodynamic properties of loop diuretics can be described by an S-shaped dose–response curve. In HF, especially during an episode of decompensation, the curve is shifted downward and to the right with reduced maximal diuretic response [3]. Beyond this dose-dependent relationship, diuretic response is dynamic during hospitalization and is subject to number of intervening factors related to the individual patient’s substrate. Thus, deriving a single, static metric to summarize this complex relationship of diuretic response is challenging. An ideal definition for diuretic response would be relatively unbiased, easily measured and practical, and utilize readily available parameters in routine clinical practice.

The American College of Cardiology Foundation/American Heart Association guidelines on the management of the hospitalized HF patient do not preference a single definition of diuretic response and broadly recommend serial evaluation of vital signs, fluid intake and output, body weights, electrolytes, renal function and signs and symptoms of congestion during active diuresis [5]. A number of more objective metrics have been proposed in monitoring response to inpatient diuresis [6]. Although a single consensus definition for diuretic response is lacking, recent clinical investigations have supported the following: change in body weight, net fluid loss or total urinary output to 40 mg of furosemide dose equivalents [7–10].

All proposed metrics of diuretic response offer attempts to quantify decongestion, but differ in subtle but important ways and may not correlate entirely with each other. For instance, there is substantial discrepancy between weight changes and net fluid loss, most likely attributable to measurement error, even in rigorously collected clinical trial data [11]. In addition, these metrics defer in how they relate to natriuresis, which more accurately reflects decongestion from the physiological perspective. Even urinary electrolyte composition changes quickly during decongestive treatment [12], suggesting that diuretic response and natriuretic response are not always interchangeable. Urine output may reflect more accurately the kidneys capacity to excrete sodium, while weight change and net fluid balance are more influenced by the baseline degree of volume overload.

A review of recent clinical data

Over the last year, four major clinical investigations, primarily based on post hoc, retrospective analyses of large randomized controlled clinical trials of hospitalized HF patients, have provided further insights into diuretic response.

Valente et al. analyzed 1745 patients with acute HF enrolled in the Placebo-Controlled Randomized Study of the Selective A(1) Adenosine Receptor Antagonist Rolofylline for Patients Hospitalized With Acute Decompensated HF and Volume Overload to Assess Treatment Effect on Congestion and Renal Function (PROTECT) trial [8]. Diuretic response was defined on hospital day 4 as change in weight (kg) per 40 mg furosemide dose equivalents. Low diuretic response was independently associated with in-hospital worsening HF, 60-day HF rehospitalization and 180-day mortality [8].

Voors et al. studied 1161 patients hospitalized for acute HF enrolled in the Efficacy and Safety of Relaxin for the Treatment of Acute HF (RELAX-AHF) trial [9]. Diuretic response was defined similar to the PROTECT analysis. Poor diuretic response was associated with less dyspnea relief through day 5 of hospitalization and higher risk of 60-day composite cardiovascular death, HF rehospitalization or renal failure, but not 180-day cardiovascular mortality [9].

Testani et al. explored diuretic response in two separate cohorts: hospitalized patients in a large academic center (n = 657) and patients enrolled in the Evaluation Study of Congestive HF and Pulmonary Artery Catheterization Effectiveness (ESCAPE) trial (n = 390) [7]. Diuretic response was defined as net fluid loss (ml) per 40 mg furosemide dose equivalents, and was independently associated with decreased post-discharge survival (more modest hazard ratio in the institutional cohort compared with ESCAPE) [7].

Most recently, ter Maaten et al. analyzed early diuretic response in 4379 patients enrolled in the Acute Study of Clinical Effectiveness of Nesiritide in Decompensated HF (ASCEND-HF) trial [10]. Diuretic response, defined as weight loss within 48 h per 40 mg of furosemide dose equivalents, was associated with worse 30-day all-cause mortality or HF rehospitalizations, but not 180-day all-cause mortality. Similar results were found using 24-h urinary volume to define diuretic response [10].

Proposed metrics: lessons & limitations

A number of common threads and lessons can be extracted from these clinical studies. First, the proposed metrics appear to be provide consistent information across studied populations. Median diuretic response was 0.4 kg weight loss/loop diuretic dose in the PROTECT, RELAX-AHF and ASCEND-HF studies [8–10], and 0.5 l net fluid loss/loop diuretic dose in Testani’s institutional study [7]. This remarkable consistency adds confidence in the reproducibility of this proposed metric. ESCAPE selectively enrolled patients with high loop diuretic requirement, and thus median diuretic response (0.15 l net fluid loss/loop diuretic dose) may be underestimated compared with a general population [7]. Second, weight or net fluid loss indexed to diuretic dose appears to provide independent, incremental information over its components. In PROTECT, the predictive ability of the diuretic response metric on post-discharge outcomes was comparable or slightly better than its components of diuretic dose or weight change [8]. Similarly, Testani et al. showed that there was only a modest correlation between diuretic response and its components [7]. Third, the determinants of diuretic response were well-established across studies and included low systemic blood pressure, elevated blood urea nitrogen, ischemic etiology of HF and diabetes [7–10]. Interestingly, the RELAX-AHF study also found regional differences in this metric with ‘Western’ countries exhibiting higher rates of poor diuretic response [9]. This may reflect regional heterogeneity in diuretic utilization patterns and comorbid disease burden. Fourth, factors classically thought to be associated with diuretic resistance may not be direct factors determining poor diuretic response. Serum creatinine and estimated glomerular filtration rate do not appear to modify the association between loop diuretic response and post-discharge events [7–9]. Rates of worsening renal function did not significantly differ by diuretic response [10]. Markers of congestion, including invasive hemodynamics [7], did not show robust association with diuretic response.

Finally, reduced diuretic response was an independent predictor of in-hospital worsening HF and post-discharge mortality and HF rehospitalization [7–10]. Data from RELAX-AHF [9] and ASCEND-HF [10] suggest that the predictive capacity of this metric is higher in the early post-discharge period, and may not extend beyond 6 months. This robust prognostic ability was replicated across multiple clinical trial populations and in a single institutional experience, was analyzed and dissected using different methodologies (median, tertiles and quintiles), and persisted after careful multivariate accounting. These studies show that diuretic response, independent of glomerular function and standard baseline characteristics, is an entity that warrants direct attention.

Important limitations to this proposed clinical definition should be considered [13]. Changes in body weight have been widely clinically adopted to determine diuretic response, but are difficult to reliably measure and are inconsistently linked with post-discharge mortality and rehospitalization [14]. Importantly, data necessary to assess these diuretic response metrics may only be available hours after diuretic dosing and as such cannot be collected real-time. Additionally, conversion of loop diuretic equivalents and true drug bioavailability of oral dosing may vary significantly from patient to patient. Finally, inpatient factors including concomitant sodium and fluid intake and use of other HF-specific agents may confound the measured diuretic response. Indeed, although the term ‘diuretic response’ has been used interchangeably for ‘loop diuretic response’ in contemporary HF literature, net diuresis per furosemide dose does not account for concomitant decongestive therapies. As with other clinical parameters and definitions, application of a single dichotomous metric to all patients hospitalized with HF (with varying clinical profiles, baseline degrees of congestion and concomitant therapies) is likely impractical. Understanding these major limitations to current definitions may help drive future efforts to apply reliable and practical metrics of diuretic response in ‘real-world’ HF populations.

Pharmacologic approach to poor diuretic response: next steps

How can we optimally approach the diuretic resistant patient in clinical and research settings? Until recently, the lack of standardized definition of poor diuretic response has hindered the development of effective strategies for this subset. As such, national committees provide limited guidance to approaching diuretic resistance [5]. Major clinical trial programs of acute HF have attempted to study patients with certain characteristics of diuretic resistant patients. Studies have selectively included patients who have high loop diuretic requirements prior to or during hospitalization [15–18]. Other studies, including Renal Optimization Strategies Evaluation – Acute HF, enrolled patients with evidence of baseline renal dysfunction (estimated glomerular filtration rate of 15–60 ml/min/1.73m²) [19]. However, these trials were not intended to specifically evaluate therapies for this distinct population.

Focusing on urinary electrolyte profile, fractional excretion of sodium and specific markers of extravascular volume status may add precision to the definition of diuretic resistance [20], but likely will detract from practicality of its use and application. Although the optimal time-point for measurement is not well-established, diuretic response measured within the first 24–48 h of hospitalization appears feasible [10] and may facilitate early use of combination decongestion therapy. A low threshold for escalation to combination diuretic therapies appeared to be a successful strategy in the control arm of the Cardiorenal Rescue Study in Acute Decompensated HF trial, even in patients with advanced cardiorenal dysfunction [21]. Once a standard definition for diuretic response has been widely validated and accepted, it may hold great promise as a key inclusion criteria in future acute HF clinical trials. This would serve to enrich the baseline risk of enrolled cohort, and identify patients who would potentially benefit from adjunctive or combination decongestive strategies. A similar paradigm has been successfully employed in the management of diuretic-refractory ascites in patients with end-stage liver disease [22]. Patients exhibiting minimal response to high-dose combination diuretics (furosemide 160 mg and spironolactone 400 mg daily) are considered for alternative interventional and device approaches. Once validated, it is also plausible that diuretic response may be utilized as a surrogate end point in future clinical trials of novel decongestive therapies.

Conclusions

Poor diuretic response has emerged as a distinct clinical entity and is associated with adverse inpatient and post-discharge outcomes in patients hospitalized for HF. It is unclear at this juncture whether diuretic resistance serves as a marker of advanced cardiorenal dysfunction or whether it represents a modifiable target for adjunctive or combination therapies. Further work is required to add precision to its definition, map its epidemiology and validate its utility in broader, ‘real-world’ populations. Once a definition has been standardized and accepted, diuretic response may represent an important inclusion criteria and end point in upcoming clinical trials in hospitalized HF to help define the optimal approach to this challenging clinical entity.

Financial & competing interests disclosure

M Vaduganathan has received research support from Novartis. AA Voors has received consultancy fees and/or research grants from: Alere, AstraZeneca, Bayer, Boehringer Ingelheim, Cardio3 Biosciences, Celladon, GSK, Johnson and Johnson, Merck/MSD, Novartis, Servier, Singulex, Sphingotec, Trevena, Vifor. J Butler has received research support from National Institutes of Health, European Union; Consultant: Amgen, Bayer, Celladon, Janssen, Novartis, Relypsa, Trevena, Z Pharma, Zensun, StealthPeptide. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.