Loop Diuretic Therapy in Left Ventricular Systolic Dysfunction: Has Familiarity Bred Contempt for a Critical but Potentially Nephrotoxic Cardio Renal Therapy?

This editorial refers to ‘The effects of adding torasemide to standard therapy on peak oxygen consumption, natriuretic peptides, and quality of life in patients with compensated left ventricular systolic dysfunction’ by S. Gupta et al., published in this issue on page 746. Long highlighted as the ‘odd man out’ in modern heart failure care, loop diuretic therapy has no cardiovascular outcomes studies for its use in any clinical setting.1 Although ostracized by some for this evidence gap, diuretic treatment remains beloved for good reason. It is often one of the first to be employed with great effect by the newly qualified physician, salvaging the critical patient from death due to hypoxic acute pulmonary oedema with rapid and gratifying effectiveness. However, has this widely recognized efficacy in this setting masked a potential for iatrogenic renal injury in less abrupt but no less fatal cardio renal decline? The use of aggressive diuretic therapy in heart failure due to left ventricular systolic dysfunction (LVSD) is well known to be associated not only with obvious benefit in acute pulmonary oedema, but it also has a linkage to progressive renal dysfunction and death in chronic LVSD.2 The practical problem is based on difficulties in the recognition of volume loading of the heart and of renal venous pressure following the relative demise of regular monitoring of invasive pulmonary pressure. Invasive monitoring has been deemed inappropriate by many critical care authorities and is itself, associated with morbidity and mortality concerns.3 As with every clinical study, the setting is the key to interpretation of therapeutic effects. Cardiorenal failure, marked by hyponatraemia; disordered electrolyte clearance; acidosis and diuretic resistance, is a prominent feature of terminal decline and death in patients with heart failure due to LVSD.4 Unlike other more dramatic events such as arrhythmic sudden death, this phenomenon has been relatively ignored in the medical literature. Loop diuretic therapy is at the heart of the management of this critical setting and may play a role in the cycle of decline given its independent but unclear linkage to cardiac death. This issue of the journal carries an excellent study5 addressing the impact of diuretic effectiveness in stable LVSD. This is of more general interest both for its emphasis on symptomatic exercise tolerance but also a range of cardiac biomarkers. The key in this study is the setting of stable patients and also of preserved renal function. The authors find little concern over the impact of the powerful loop diuretic torasemide, in what is presumed to be largely euvolaemic LVSD. However, volume status is rarely easily defined in either stable or symptomatically decompensated patients6 (often but certainly not always hospitalized with acute pulmonary oedema). That there is no impact in this experiment on cardiac performance and exercise capacity is perhaps therefore of little surprise. With baseline volume status in this study basically unclear, a proportion may have been slightly overloaded whereas another sample may have been relatively dehydrated. Should this make us relaxed about escalating diuretics regardless of accurate assessment of volume status? Progressive diuretic dose escalation in response to persistent or worsening symptoms is common.7 However, it may not only fail to overcome ‘apparent diuretic resistance’ (here defined loosely as a failure to improve symptoms) it can exacerbate poor glomerular filling, reducing drug filtration, and delivery into the intra tubular filtrate where its mechanism of action is sited. This cycle can lead to progressive acute renal deterioration and systemic arterial hypotension which can be fatal. Moreover the solution to this cycle may involve temporary drug withdrawal [often of non-diuretic strategies such as Renin–Angiotensin–Aldosterone system (RAAS) inhibitors] and/or invasive pressure monitoring. Both of these techniques have experienced a backlash in recent times. The former strategy is rarely reported in practice, and many will note considerable variations in practice between clinicians in your own institution, and the latter is now relatively infrequently performed. Invasive pressure monitoring, while an accepted gold standard, might be seen as too ‘aggressive’ for many of the very elderly subjects commonly presenting with heart failure syndromes. There are few surrogates for an accurate definition of volume loading status of not only left heart function but perhaps more cogently renal venous pressure (responsible for post renal resistance).8 Both are now rarely assessed fully in routine practice. Escalating diuretic therapy is by far and away the standard clinical response to symptom recurrence. The proportion of symptomatic patients who have genuine central cardiac volume overload is often unclear. There is no accepted surrogate for estimating central volume loading short of invasive pressure monitoring. Conventional clinical techniques such as symptoms, physical signs (with some exceptions),6,9 and simple tests such as chest radiographs and body weight measures are notoriously insensitive and non-specific to the point of being misleading. The key factor in this is not only reliably defining volume loading of the heart/pulmonary pressure but more significantly an integrated renal glomerular and tubular perfusion. Diuretic resistance is complex10 with multiple potential mechanisms in individual patients. The concern here is the hypovolaemic LVSD patient where failure of diuretic response (symptomatic or otherwise) is responded to by diuretic escalation. Both pharmacokinetic and pharmacodynamic changes can occur that can impact on diuretic filtration at the glomerulus. Volume filling and forward cardiac output is just as critical to the renal glomerulus as the pulmonary bed. The former often has much less room for manoeuvre in treated LVSD given the routine (if often incomplete) paralysis by drug therapy of key homeostatic mechanisms such as the sympathetic nervous system; RAAS and aldosterone. This therapeutic paralysis combined with borderline arterial pressure is a heady mixture with potential for iatrogenic renal injury. Although this is based partly on deteriorating ventricular dysfunction it may also be due to what can be seen by some as a huge pressure for dose escalating polypharmacy based on multicentre trials and protocolized care.11 The answer to some cases of intravascular hypovolaemia may be in suspending neurohormonal suppression rather than progressive diuretic dose escalation. How do we define that? The vast majority of patients are managed by symptoms, signs, and basic tests alone, however, many cardiologists and nurses also focus on cardiac biomarkers of response to justify therapeutic changes. Using blood brain natriuretic peptide and derivatives has excellent sensitivity and some utility provided that their lack of specificity is factored in.12 Many practitioners are willing to accept symptomatic worsening, radiographic images or non-specific markers, and signs such as body weight and ankle swelling as justification for hospitalization for aggressive combination loop diuretic therapy and dose escalation to treat presumed ‘diuretic resistance’. Unfortunately these routine non-specific assessments provide imprecise and inadequate estimates of cardiac or renal volume loading. In patients already affected by intravascular dehydration (often with chronic extravasation in non-critical beds, e.g. ankle swelling), this path can steeply accelerate cardiorenal decline and hasten death. There is no doubt that progressive escalation of diuretic therapy can lead to a state of relative diuretic resistance through eventual cessation of glomerular filtration and falling renal perfusion.13 This phenomenon is about both renal and cardiac perfusion and performance. The key may lie in within subject repeated measures and both renal or cardiac biomarkers or physical biomeasurements (such as bioimpedance) may be relevant in combination with indices of cardiac function.14 Modern pacemakers and defibrillators can incorporate forms of repeated bioimpedance measurement.15 Intrathoracic impedance, reflecting pulmonary congestion, is a complex measurement that may not yet be optimized in implantable systems, but at least can create a running average signal whose trend can suggest altered whole body water.16 Unfortunately such physical measures are not always specific for fluid shifts and altered diuretic therapy in response to these can create worsening hypotensive symptoms and injury. More device manufacturers are incorporating this technology into both CRT and pacemaker systems yet its sensitivity and specificity between or within any of the huge range of subjects that might receive them (many without LVSD), are not well-defined. What other factors raise bioimpedance unrelated to intravascular fluid volume or how to integrate this with clinical assessments and either cardiac or renal biomarkers is unclear. Is it feasible to define net glomerulo-tubular perfusion non-invasively? The sensitivity and specificity of conventional renal measures (estimates of GFR or tubular injury) is untested in patients with multiple neurohormonal suppression. Although most heart failure teams see cases of catastrophic hypovolaemic acute renal failure, more subtle and gradual onset hypovolaemia (particularly that occurring in conjunction with chronic and fixed peripheral extravasation) is very hard to define. We do need to explore new ways of identifying this process outside conventional measures and before injury is quite so dramatic (and potentially irreversible). The challenge here lies not just in an examination of the safety of loop diuretic in asymptomatic LVSD patients.5 This is an interesting experiment but probably not overly sensitive enough for occult renal injury, and not carried out in those with deteriorating symptoms, which is when physicians most frequently augment therapy. We must use this to look in a more integrated fashion at markers of volume loading not only of the heart but of the kidneys. Strategies need to be developed in practical terms to help identify patients where diuretic based dose escalation protocols need to be curtailed. Perhaps re-examination of the current reluctance to conduct invasive cardiac and right heart pressure monitoring is required. New renal biomarker technologies17 need to be explored and investigated in directing routine volume manipulation through diuresis; off loading; inotropic or renal replacement therapies. Critically a change in the clinical setting of individual patients needs to better understood. The evolving impact of progressive changes in LVSD care based on protocols has some risk where these facilitate escalating therapy in response to non-specific or even static symptoms. Not only is this unlikely to be effective in every case, it may also set up a path for decline if applied blindly without recognition of the subtlety of cardio renal perfusion. Protocols and individual care plans should include provision for reduction of diuretic doses in those where euvolaemia is achieved and especially when intravascular hypovolaemia occurs, either chronically or acutely, e.g. in response to gastrointestinal disturbance. It is time to take the lead from these experiments and re-examine the interdependence of the kidney and heart specifically with respect to the optimization of volume loading not only of the heart but the kidneys. Multicentre clinical trials have told us a lot about the dosing of additive therapy and their impact on deferring events. Although most patients with LVSD have immutable mortality, all physicians have a duty to avoid causing iatrogenic injury. Cardiac physicians perhaps need to better understand the insensitivity and non-specificity of chronic exercise limitation and be more alert to detecting declining function in organs other than the heart. Given the critical role of combination heart failure therapy with neurohormonal antagonists in this setting, a new set of LVSD trials need to explore the validity and safety of continuing these at high dosage along with assessing the impact of augmenting loop diuretic. These studies could explore both physical (including invasive haemodynamics) and blood biomarker panels not only directed at the heart but at key organs such as the kidneys. We should also formally explore the forgotten skills of down titration of chronic treatments, temporarily or permanently, in a timely fashion. Perhaps too few of us treating these patients on a daily basis seem to understand the impact of doing otherwise. Conflict of interest: none declared.