An adequate response of the kidney to changes in perfusion pressure and volume status is a central factor in the normal homeostasis of circulating volume that involves a coordinated response of the vascular and tubular part of the nephron. Under normal circumstances, the auto-regulatory capacity of the glomerular microcirculation allows filtration capacity to remain constant over a wide range of pressures, thus ensuring stable and continuous excretion of metabolic waste products despite changes in volume status and blood pressure. Meanwhile, the tubular component of the homeostatic response ensures stability of sodium and volume balance by adapting proximal and distal tubular sodium reabsorption. These adaptive glomerular and tubular changes occur in response to altered renal perfusion pressure, as well as in response to changes in volume status without changes in renal perfusion pressure. Thus, renal adaptive capacity allows the body to maintain circulatory homeostasis over a widely different range of sodium and fluid intake and provides our evolutionary defense against volume depletion by blood loss and dehydration.

During impairment of renal perfusion pressure, glomerular filtration rate (GFR) is maintained by a two-step hemodynamic adaption. First, pre-glomerular (afferent) vasodilatation occurs by the afferent myogenic response, reflecting the classical auto-regulation response of the kidney that serves to maintain both glomerular flow and pressure, and hence filtration, with an additional regulation of afferent tone by the tubuloglomerular feedback loop. Briefly, the latter mechanism works as follows: renal perfusion decline leads to compensatory tubular sodium reabsorption and therefore to a reduction in distal tubular sodium delivery, which is sensed by the macula densa. Lower distal sodium delivery decreases adenosine concentration and therefore adenosine-α1 receptor activity, causing afferent vasoconstriction, commonly known as tubuloglomerular feedback. When faced with a further fall in perfusion pressure, post-glomerular (efferent) vasoconstriction occurs as the next resort of the auto-regulatory response. This ensures preservation of filtration pressure in glomerular capillaries at the expense of a further fall in perfusion. Thus, the proportion of renal perfusion effectively filtered, the filtration fraction (FF) is increased, as a distinguishing feature of increased efferent vascular tone. Increased activity of the renin–angiotensin–aldosterone system (RAAS), elicited by the decreased renal perfusion pressure, plays an important role in the renal hemodynamic and tubular response to perfusion impairment. First, the efferent vasoconstriction is mainly mediated by angiotensin II. Concomitant intra-renal production of vasodilator prostaglandins precludes inadvertent vasoconstrictor effects of angiotensin II on the afferent arteriole that could threaten filtration pressure under these circumstances. Second, the concomitant increase in tubular sodium reabsorption is mediated by angiotensin II, by direct tubular effects as well as by an increase in aldosterone. The ensuing renal sodium retention is facilitated by the concomitant decrease in natriuretic peptides. Angiotensin II furthermore contributes to maintenance of blood pressure by its vasoconstrictor effects on the systemic vascular bed.

In chronic heart failure, whether with reduced or preserved ejection fraction, considerable impairment of renal perfusion occurs due to the decrease in cardiac output. Even if cardiac index is only mildly impaired, the perfusion impairment can be considerable [3, 4]. The reduction in renal perfusion is the main determinant of the impairment in renal function that is usually observed in heart failure [5]. It would be logical to assume that the reduction in renal perfusion and renal function is proportional to the decrease in cardiac output. However, the association between cardiac index and renal function has been notoriously difficult to demonstrate. In several studies, improvement in cardiac index or improvement in pulmonary wedge pressure did not adequately predict improvement in renal function in heart failure patients [6‐9]. To understand this seeming discrepancy, it is important to realize that the association between cardiac output and impairment of GFR is not mono-dimensional.

First, renal perfusion and hence GFR can be affected by pre-existing renal arteriolosclerosis. Second, the renal auto-regulatory capacity can maintain GFR in spite of renal perfusion impairment, by tubuloglomerular feedback and afferent vasodilatation, and by increased RAAS activity and consequent efferent vasoconstriction. This adaptive mechanism is readily apparent when renal perfusion is measured concomitantly with GFR, from the increase in FF that is usually present in patients with chronic heart failure. Accordingly, when only data on (estimated) GFR are available, the association between cardiac index and renal changes is obscured. It is only in advanced heart failure, when renal perfusion is severely impaired, that FF decreases sharply, probably because renal perfusion is maintained at the expense of locomotor muscle perfusion initially [10] and not until heart failure advances, will renal perfusion decrease below the lower threshold of auto-regulatory capacity [5]. Renal auto-regulation itself may become impaired due to decreased availability of nitric oxide [11], which impairs tubuloglomerular feedback. Decreased nitric oxide availability, caused by endothelial dysfunction, is not only seen in heart failure but also in diabetes mellitus, inflammation, and atherosclerosis. In such a situation, GFR becomes strongly dependent on blood pressure that is, understandably, very low as well with consequently a sharp further decrease in GFR [5]. It should be noted that currently most patients with chronic heart failure are treated by RAAS blockade for its cardio-protective effects. RAAS blockade, however, impairs the efferent contribution to auto-regulatory capacity and thus can aggravate the GFR impairment. The latter can, to a certain extent, be considered to be a price to be paid for the cardio-protective effects of the RAAS blockade.

The above described sequence of events represents the pre-renal component of the GFR impairment in chronic heart failure and corresponds to the renal changes that occur in other pre-renal conditions, such as severe dehydration and volume depletion. Accordingly, when renal perfusion is restored by volume repletion, renal function will improve. The major difference between pre-renal failure due to pure volume deficit and renal function impairment in chronic heart failure is, obviously, the congestive state. In patients with pure volume deficit, volume repletion is the single effective treatment, whereas in chronic heart failure, volume repletion measures can be expected to worsen the congestive state.

In the clinical management of heart failure, therefore, the usual concept on volume management is that of a trade-off: measures to improve congestion, diuretics and dietary sodium restriction will do so at the expense of a further decrease in renal function. Vice versa, when severe renal function impairment prompts for accepting a positive volume balance by reducing diuretics or a more liberal sodium intake, the improvement in renal function occurs at the expense of worsening congestion.

Until recently, it was assumed that the congestive state as such did not impact on renal function, despite a few very old studies [12‐14] that suggested an adverse effect of high venous pressure on renal function. Several studies, however, recently demonstrated an association between venous congestion and worse renal function, in various populations with cardiac impairment from different origin with both reduced and preserved ejection fraction.

A large retrospective analysis in 2557 patient from our center that underwent right heart catheterization showed that central venous pressure (CVP) is associated with estimated GFR (eGFR) in a bimodal fashion, with the highest eGFR at a CVP of approximately 3 mmHg and a progressively lower eGFR in subjects with CVP above 5–6 mm Hg, comprising approximately one-third of the population. Remarkably, is shown in Fig. 1, this bimodal association was independent of cardiac index, demonstrating that the adverse impact of high CVP on eGFR was not merely due to the patients with high CVP being a subset in a very poor condition including worse cardiac output. In fact, the decrease in eGFR at higher CVP was, if anything, steeper in subjects with a relatively preserved eGFR. Elevated CVP was also an independent predictor for mortality [15]. Not only directly measured CVP was found to be related to renal function. In a separate study in 2647 patients with systolic heart failure, presence of congestive symptoms, such as peripheral edema, elevated jugular venous pressure, orthopnea, and ascites were associated with worse eGFR and, moreover, with mortality [16].

The separate contributions of renal perfusion impairment and congestion were analyzed in a population of patients with predominantly right-sided heart failure due to primary or secondary pulmonary hypertension, screened for lung transplantation at our center. Their work-up included renal hemodynamic measurements as well as right-sided heart catheterization. In this population with a mean true GFR of 73 ml/min/1.73m2, renal blood flow and right atrial pressure were the only determinants of GFR on multivariate analysis, with a worse GFR in subjects with lower renal blood flow and higher right atrial pressure. Remarkably, elevated right atrial pressure did not affect GFR in subjects with preserved renal blood flow, whereas it was associated with a significant further reduction in GFR in subjects in whom renal blood flow was impaired (Fig. 2) [17]. This leads to the concept that venous congestion or backward failure impairs GFR only, or preferentially, when forward failure is concomitantly present.

The mechanisms of the adverse effect of high venous pressure on renal function are incompletely elucidated. Animal studies, in a canine model, where induction of volume overload was associated with a rise in renal interstitial pressure along with a decline in renal perfusion and renal function, suggest that elevated venous pressure may translate into inappropriate elevation of renal interstitial pressure as a possible mechanism [18].

In addition, it has been shown that venous congestion and associated endothelial stretch increase the production of pro-inflammatory cytokines [19], such as tissue necrosis factor-α, which are associated with sodium retention, renal hypertrophy, and nephropathy [20, 21]. Additional production of reactive oxygen species may reduce nitric oxide availability, leading to peripheral and systemic vasoconstriction [22]. This does not only increase cardiac filling pressure via increased central blood volume but also reduces renal perfusion.

The evidence that congestion as such is associated with an adverse effect on renal function challenges the trade-off concept that volume repletion, with acceptance of a certain worsening of congestion, will invariably improve renal function. In fact, the curve in Fig. 1 suggests that careful targeting of the congestive state can lead to improvement in renal function in subjects with severely elevated CVP, which would be in line with the aforementioned animal study. It should be noted here that the associations between CVP and eGFR in human populations were cross-sectional and did not include interventions in volume status. In fact, recent data on intervention in volume status in heart failure are scarce.

Targeting volume status can be effectuated by dietary sodium restriction, diuretics, and the combination, whereas in severe heart failure, hyponatremia can prompt for fluid restriction as well. Ultrafiltration can be used as a last resort in patients with severe congestion along with poor cardiac output.

The ESC guidelines on acute and chronic heart failure [23] identify the effect of dietary sodium restriction as one of the gaps in the current evidence. Studies on natriuretic efficacy of diuretics, however, have shown that negative volume balance is difficult to achieve by diuretics when sodium intake remains high, especially when the kidney is avidly retaining sodium [24]. It seems logical therefore to aim for dietary sodium restriction in patients where diuretic treatment is considered to be indicated. This requires that patients should be educated concerning the salt content of food, as recommended by the guidelines. The latter is important as, first, salt content of the western diet (9–12 g/day) contains approximately double the amount recommended for the general population, i.e., 5–6 g/day, and moreover, 75–80% of dietary sodium intake derives from pre-manufactured food products. In the clinical management of heart failure, the actual adherence to dietary sodium restriction is rarely verified, but it is likely that many patients ingest excess sodium without being aware of it, as has been shown in studies in chronic kidney disease, where dietary sodium intake, as apparent from 24-hour sodium excretion, is approximately similar to the general population, despite efforts to reduce salt intake [25‐27].

Use of diuretics is, in different guidelines [23, 28], limited to patients with symptoms of congestion. This conservative use of diuretics is mainly based on two concerns. First, during therapy with diuretics, GFR tends to decline [29], and reduction of GFR in general is a poor prognostic sign. Second, diuretics induce excess neurohormonal activation, in particular of plasma renin activity and aldosterone [30], which could adversely affect outcome. However, in the current era, with most patients being treated with RAAS blockade, often in combination with beta-blockade, the significance of the latter may have diminished [23, 28], although concern remains on the direct pro-fibrotic effects of aldosterone on the heart and the kidney [31]. Aldosterone antagonists therefore look promising as reno-protective agents [32], but newer B-blockers may also yield protective effects on renal function [33]. These drugs act via an effect on intrarenal vascular resistance and may therefore affect renal blood flow and thus GFR. Larger trials in heart failure are needed to study their effect on renal function in heart failure.

Recent data, however, warrant a re-appraisal of the impact of volume targeting on GFR in patients with heart failure. It should be noted first that a diuretic-induced decrease in GFR in patients on RAAS blockade reflects reversible renal hemodynamic changes rather than persistent renal damage. This is probably due to the impaired autoregulation under these conditions. Withdrawal of diuretics or restoration of volume status by a liberal sodium intake will restore GFR, as demonstrated in renal patients (Fig. 3). Interestingly, in these renal patients, the treatment-induced decrease in GFR predicts a favorable long-term renal outcome [34], probably because it reflects the efficacy of RAAS blockade apparent from the consequent decrease in filtration pressure. A corresponding finding was recently reported from a well-performed study in decompensated heart failure patients [35], where subjects aggressively treated with diuretics had worse renal function, but 180-day mortality was reduced in these subjects, when compared to subjects treated more conservatively. Worsening of renal function may therefore be accepted upon start of therapy with diuretics or sodium restriction (Fig. 3).

Even more recently, results from the DOSE trial were released, wherein subjects admitted to the hospital with acute decompensated heart failure were treated with either diuretics continuously or via bolus and either normal or high doses. Subjects with higher doses of diuretics fared better subjectively, which was associated with more weight loss and greater urinary output. The greater rise of serum creatinine in this group was completely reversible after 3 weeks [36].

Therefore, data on (estimated) renal function in a decompensated state should be interpreted with caution as worsening GFR may reflect both poor prognosis and proper treatment response, which is associated with good prognosis. Without information on filtration fraction or renal perfusion, which is usually not readily available, eGFR and serum creatinine are therefore poor prognostic markers of renal function during acute decompensated heart failure. Due to these reasons, it has been suggested that blood urea nitrogen/creatinine ratio may be a better prognostic marker in ADHF than serum creatinine or estimated GFR [37].

Studies in hypertension and chronic kidney disease provide clues that suggest proper attention for volume targeting will be beneficial in heart failure. This not only applies to amelioration of the congestive state and hypertension but notably also to the potentiation of the effect of RAAS blockade in those conditions [38, 39], with concomitant beneficial effects on hard end points [40]. Considering the main role of RAAS blockade in the clinical management of heart failure, it might be worthwhile to consider a systematic re-appraisal of the role of dietary sodium restriction and diuretics in the management of heart failure, especially in view of the fact that counseling on sodium restriction for a short period may bear significant long-term effects in terms of reductions in cardiovascular mortality [40]. The strategy for volume targeting in heart failure will usually require both dietary sodium restriction and diuretics to be sufficiently effective. As noted above, diuretics are insufficiently effective in the presence of high sodium intake [24]. Whereas the effects of low sodium diet and diuretics exert more or less similar effects on volume status, it should be noted that they may differ in other effects relevant to heart failure patients, such as for instance serum potassium. Of note, in renal patients, recent data showed that diuretic treatment affects EPO production when used in conjunction with RAAS blockade [41] in spite of a well-preserved renal function, whereas low sodium diet does not (Fig. 4). Considering the prognostic impact of EPO levels in heart failure, and the corresponding treatment regimen, this might bear relevance to heart failure patients as well. Corresponding data for heart failure patients are not available yet, but the data in renal patients suggest that it would be worthwhile to explore not only the effects of volume targeting but also to address whether it matters for long-term outcome by which strategy the volume correction is achieved.