Combined Neprilysin and RAS Inhibition for the Failing Heart: Straining the kidney to Help the Heart?

This article refers to ‘Renal effects of the angiotensin receptor neprilysin inhibitor LCZ696 in patients with heart failure and preserved ejection fraction’† by Adriaan A. Voors et al., published in this issue, European Journal of Heart Failure (2015); 17: 510–517. Atrial, brain and c-type natriuretic peptides (NPs) belong to a family of hormones secreted from atria and the left ventricle in response to excess plasma volume and left ventricular filling pressures. In patients with heart failure (HF), they help maintain sodium and fluid balance by reducing renal and systemic vascular resistances and facilitating natriuresis and diuresis despite excess renin–angiotensin–system (RAS) activation.1 As the neutral endopeptidase neprilysin contributes to the breakdown of NPs and other endogenous (vasodilatory) peptides, including bradykinin, neprilysin inhibitors soon emerged as an adjunct therapy to angiotensin-converting enzyme (ACE) inhibitors to alleviate fluid overload and improve systemic haemodynamics in HF patients by increasing endogenous NPs bioavailability.2 Indeed, omapatrilat, the progenitor of dual-acting neprilysin and ACE inhibitors, reduced mortality and hospitalizations more effectively than enalapril in HF patients with reduced ejection fraction.3 The medication, however, was not approved by the Food and Drug Administration and was withdrawn from development because of an unacceptable excess risk of even fatal angioneurotic oedema associated with increased bradykinin levels. Thus dual-acting angiotensin-receptor–neprilysin inhibitors were developed to combine the benefits of combined RAS and neprilysin inhibition without excess risk of angioedema. Indeed, unlike ACE inhibitors, angiotensin receptor blockers (ARBs) have minimal effect on bradykinin activity and, in combination with neprilysin inhibitors, are much less likely to increase bradykinin bioavailability. LCZ696, the ancestor of these compounds, combines, in a 1:1 molar complex, the ARB valsartan with AHU377, a prodrug that, following oral administration, is rapidly converted into the active neprilysin inhibitor LBQ657.4 The possible benefits of LCZ696 over ACE inhibitor monotherapy were formally tested in 8442 HF patients in New York Heart Association (NYHA) class II, III or IV with an ejection fraction of ≤40%, who were included in the PARADIGM-HF trial.5 The study was prematurely stopped because the predefined boundary for an overwhelming benefit with LCZ696 on the composite endpoint of cardiovascular deaths and hospitalizations for HF was crossed. At final analyses, LCZ696 was associated with significantly fewer all-cause and cardiovascular deaths, HF hospitalizations, hyperkalaemia and cough compared with enalapril. These benefits largely exceeded the higher incidence of hypotension and non-serious angioedema observed with LCZ696.5 Consistently, the PARAMOUNT study showed that 36-week treatment with LCZ696 reduced N-terminal pro-brain natriuretic peptide (NT-proBNP; a marker of left ventricular wall stress) more effectively than valsartan and improved NYHA class, blood pressure (BP) and left atrial size in 301 HF patients with preserved ejection fraction.6 The drug was well tolerated, and although one case of angioedema did occur, it did not require hospitalization. However, the study was underpowered and with too short a follow-up to assess the protective effects of LCZ696 against cardiovascular events in this context. In addition to improving systemic haemodynamics, increased NP bioavailability by neprilysin inhibition is expected to reduce the risk of worsening kidney function, a relatively frequent event in HF patients exposed to ACE inhibitors or ARBs. However, to unravel the specific renal effects of LCZ696, in particular of the LBQ657 moiety, and RAS inhibitors in this context we need to go back to the old days of Arthur J. Merril who, in 1946, measured renal plasma flow (RPF) and glomerular filtration rate (GFR) by using the sodium para-amino hippurate and inulin renal clearance techniques, in 37 subjects with HF from different aetiologies.7. In these patients the RPF was reduced from one-third to one-fifth of what is normal, whereas cardiac output was rarely reduced below one-half the resting value. Thus, in these patients, there was a specific diversion of blood away from the kidneys, organs which normally receive about 20% of the cardiac output. Despite the large reduction in RPF, however, the GFR was one-half to one-third of normal. This was explained by a concomitant increase in the filtration fraction, which, according to Merrill, reflected ‘high intraglomerular pressure from efferent arteriolar constriction, sustained by enhanced renal renin release’. This was the first demonstration that, in stable HF, reduced kidney perfusion caused by circulatory impairment leads to a compensatory increase in the filtration fraction, which preserves intracapillary pressure and the GFR despite decreased perfusion pressure (Figure 1, Panel a). This adaptive mechanism is blunted by RAS inhibitors that, by preventing angiotensin II-mediated autoregulation, render intraglomerular hydraulic pressure and GFR directly dependent on perfusion pressure and any treatment-induced BP reduction may result in appreciable increases in serum creatinine levels.8 However, if the ACE inhibitor dose is titrated to BP response and persistent, severe hypotension is avoided, the above haemodynamic phenomena seldom translate into progressive renal function loss and, based on observations in other clinical settings, might even be renoprotective.9 Interestingly, in the PARAMOUNT renal substudy reported by Voors et al. in this issue of the Journal,10 36 weeks of valsartan treatment induced a decrease in estimated GFR that was not observed in the LCZ696 treatment group. Although these data must be taken with caution because of the limitations of serum creatinine-based formulas in GFR estimation,11 the worsening in kidney function associated with valsartan could be explained by treatment-induced reduction in systemic BP and concomitant loss of renal auto-regulation (Figure 1, Panel b). Conversely, in the LCZ696 treatment group the haemodynamic changes induced by valsartan were most likely prevented because of the enhanced bioavailability of NPs achieved by neprilysin inhibition. Indeed, these peptides may directly increase intraglomerular capillary pressure by preferentially dilating the afferent arteriole and relatively constricting the efferent arteriole. Conceivably, these haemodynamic effects overcome those of angiotensin inhibition and enable the maintenance of GFR even despite a substantial reduction in systemic BP (Figure 1, Panel c). This hypothesis needs confirmation in studies directly measuring kidney perfusion and filtration, as treatment-induced changes in renal haemodynamics might affect long-term kidney function outcome independent of the progression of HF and associated comorbidities.12 Indeed, ever since investigators from the Brenner group in Boston first reported that intrarenal actions of NPs contribute significantly to the pathogenesis of glomerular hyperfiltration in diabetic rats,13 evidence has emerged that the renal haemodynamic changes sustained by increased NP levels may contribute to the onset and progression of renal disease, both in experimental animals and in humans.14 Thus, the association of LCZ696 with a statistically significant and clinically relevant increase in urinary albumin to creatinine ratio, was not surprising.10 This finding must be carefully considered as albuminuria, even in the normal range, is an independent renal and cardiovascular risk factor and, independent of the underlying causes, any increase in urinary albumin might translate into an excess risk of renal and cardiovascular events,15 even in subjects with heart failure.16 The increase in albuminuria was observed despite a substantial reduction in BP, which, per se, would be expected to decrease urinary albumin. Thus, the finding necessarily reflected a specific renal effect of LCZ696, most likely mediated by increased bioavailability of NP. This hypothesis is consistent with previous evidence that the infusion of atrial natriuretic peptide (ANP) may increase albuminuria in healthy volunteers and to a greater extent in patients with essential hypertension or type 1 diabetes and microalbuminuria. Even more prominent effects were reported in patients with non-diabetic renal disease and a nephrotic syndrome. In all the above studies, ANP infusion was considered to induce or amplify a defect in the size-selectivity of the glomerular barrier because of increased intracapillary pressure and concomitant relaxation of mesangial cells with a secondary increase in glomerular filtering surface area. Thus, the impaired sieving function resulted in enhanced ultrafiltration of circulating albumin and even larger proteins and circulating macromolecules.17 Increased protein ultrafiltration and tubular overload would eventually initiate and sustain progressive renal damage through a series of mechanisms, including chronic inflammation, ischaemia, oxidative stress and others.18 Unfortunately, the PARAMOUNT trial was largely underpowered and its follow-up was too short to assess whether, and to what extent, the enhanced albumin excretion induced by LCZ696 might herald and sustain the onset and progression of renal disease in HF subjects with preserved ejection fraction. Direct measurements of hydraulic intraglomerular pressure are impossible in humans, but—as demonstrated by Merril and co-workers7—renal haemodynamics can be evaluated precisely by measuring the clearance of exogenous markers of kidney perfusion and filtration. Calculating the filtration fraction provides an indirect estimate of intracapillary pressure, and the measurement of the fractional clearance of endogenous macromolecules such as albumin and IgG and, ideally, of exogenous polysaccharides of different radius makes it possible to assess whether, and to what extent, increased albuminuria is sustained by intraglomerular hypertension and/or directly impaired glomerular barrier sieving function. The possibility of a concomitant NP-induced impairment of tubular handling of ultrafiltered proteins should also be taken into consideration by measuring the urinary excretion of β2-microglobulin and free κ-light chains.14 These studies will make it possible to assess whether the haemodynamic effects mediated by neprilysin-induced enhanced NP bioavailability—that by no means improve the performance of failing hearts in the short-term—might induce and accelerate progressive renal function loss in the long run. Worsening kidney function would translate into increased preload and afterload to the failing myocardium which, in combination with other metabolic and functional abnormalities of uraemia, would accelerate rather than prevent the progression of HF in a vicious circle that would be difficult to break. Thus, adequately designed studies evaluating glomerular perfusion, filtration, and sieving function by gold standard techniques (7) are needed to unravel the mechanisms underlying the renal effects of LCZ696 in HF patients, and to assess whether and to what extent these effects may differ between those with or without impaired ejection fraction. This might make it possible to identify parameters (such as GFR and albuminuria) that might help tailor valsartan and LBQ657 dosing, and possibly timing and duration of intervention, to individual patient response This could be instrumental to optimizing the benefits of combined RAS and neprilysin inhibition on systemic haemodynamics, and at the same time minimizing the stress imposed by NPs on the microvasculature of the kidney. However, a randomized clinical trial specifically addressing the long-term renal effects of LCZ696 will require thousands of patients to be followed for years, and might be confounded by the competitive risk between renal and cardiovascular events.