29 Renal dysfunction

Up to 55% of patients with heart failure (HF) have evidence of chronic kidney disease (CKD) stages 3–5 (estimated glomerular filtration rate (eGFR) of 15–59 mL/min, see Fig. 29.1), and mortality rises in proportion to fall in GFR (see Fig. 29.2).¹ In such patients, advanced CKD is as prognostically important as left ventricular ejection fraction (LVEF).² HF and its treatment may also play an important role in the pathophysiology of acute kidney injury (AKI), with a further associated risk of adverse outcome. Cardiovascular disease is the leading cause of death in patients with CKD, and structural cardiac abnormalities are highly prevalent in dialysis patients.

Mortality in patients who have both renal and cardiovascular disease is very much higher than in the general population. The difficulty in defining a precise epidemiological association between HF and CKD is that much of the available data are derived from clinical trials and studies that have strict inclusion/exclusion criteria. These selection criteria limit the understanding of how renal disease impacts on other medical conditions, not just HF, as patients with advanced CKD in particular are typically excluded from clinical trials. However, the majority of elderly patients with HF will have some degree of CKD, as may many younger patients. A prospective cohort study of all comers to a HF clinic found that less than 17% of patients had a normal creatinine clearance (which was the previously used surrogate for eGFR).² The presence of renal impairment in patients with HF confers a major detrimental impact upon survival. For patients with advanced CKD, mortality increases by 1% for each 1 mL/min fall in creatinine clearance.²

HF is prone to develop and progress in patients of all ages with endstage renal disease (ESRD). Many factors contribute, including hypertension (found in 〉90% of patients with ESRD), anaemia, and fluid overload. 74% of patients have echocardiographic evidence of left ventricular hypertrophy (LVH) when starting renal replacement therapy (RRT), 36% left ventricular dilatation, 15% severe left ventricular dysfunction; and 4.5% of the patients having dialysis fulfil ACC/ASA/ESC criteria for ICD implantation based on primary and secondary prevention studies of patients with HF.^3,–5 Foley et al.⁶ followed a cohort of 259 patients from the time of starting dialysis for a mean of 41 months, and assessed baseline and follow-up echocardiography. In this study 70% of patients had an increase in left ventricular mass index (LVMI) and 50% an increase in left ventricular cavity volume at the end of the study compared to baseline values; 33% of patients developed HF, one-half of which were de novo episodes. Furthermore, each 10-mmHg rise in mean arterial pressure was associated with a relative risk of de novo HF of 1.44.

Table 29.2 shows the relative annual mortality figures for patients with anaemia, CKD and HF, or with combinations of these conditions⁷ in a study of a random cohort of 5% of Medicare database patients (1 321 156 subjects). The patients were subdivided according to the presence or absence of anaemia, CKD (excluding ESRD) and HF, identified as comorbidities on Medicare claims. The relative risk of death in the presence of these diseases, matched for age and other comorbidities against the remainder of the cohort, is shown. The annual mortality was 4% for patients with no history of anaemia, CKD, or HF; 8% for patients with CKD; and 23% for patients with all three comorbidities.

As a general rule, all-cause mortality rises significantly as GFR falls. The excess mortality includes a disproportionate number of deaths due to left ventricular pump failure⁸ and the mortality risk is equally high in patients with both systolic and diastolic dysfunction.⁹ Data from the United States Renal Data System shows that cardiomyopathy, congestive HF, or pulmonary oedema is the primary cause of mortality in dialysis patients with a rate of 11.4 events per 1000 patient years.¹⁰

The term ‘cardiorenal syndrome’ has been used to describe the common finding of AKI in patients admitted to hospital with decompensated HF. The term has also been used to describe a worsening of renal function in response to HF treatment, and the barrier to management that it may cause. Ronco et al.¹¹ classify the interaction of chronic as well as acute cardiorenal disease into five types of cardiorenal syndrome (Table 29.3). Decompensated HF as a cause or effect of AKI is type 1 or 3 respectively, CHF as a cause and effect of CKD is type 2 or 4 respectively. Type 5 is classed as cardiovascular and renal end-organ damage from a common underlying pathology, most often diabetes mellitus, atherosclerosis, and/or hypertension. Though categorized as separate clinical entities, each of these types of cardiorenal syndrome is more likely in the context of another. For example, AKI is more likely in a patient with decompensated HF if the patient has underlying diabetes, associated CKD, and coronary artery disease (CAD) than if the patient does not have such a history. What is more, because dysfunction of one organ system may cause or exacerbate dysfunction of the other, a ‘snowball’ effect can occur. This in part explains the poor outcome in patients with both heart and kidney failure.

Cardiovascular and renal disease may be due to the same underlying disease. Furthermore, CKD is an independent risk factor for developing cardiovascular disease, particularly CAD and LVH. CAD is responsible for more than one-half of incident cases of HF,¹² and arterial disease is a major cause of renal disease.

Hypertension is extremely common in patients with CKD and it is associated with a high rate of de novo cardiac failure and ischaemic heart disease, especially in those on dialysis. Even modestly elevated blood pressure is associated with LVH and cardiomyopathy. However, in the study of dialysis patients by Foley et al.⁶ discussed above, only a mean arterial pressure of 106 mmHg or more was independently associated with de novo HF—tight control of hypertension is fundamental to the renal physician’s practice. However, in ESRD mortality is most strongly associated with a low blood pressure⁶ and pump failure will lead to a low blood pressure irrespective of renal function. Hence, if a dialysis patient develops significant HF, their requirement for antihypertensive therapy may need to be re-evaluated.

Three-quarters of all dialysis patients have echocardiographic evidence of LVH. The presence of LVH in ESRD is associated with an adjusted relative risk of cardiac death of 2.7 (95% CI 0.9–8.2).¹³ LVH develops early in CKD; in a study of 175 consecutive patients attending a predialysis clinic, LVMI increased as creatinine clearance fell and LVH was independently associated with age and systolic hypertension.¹⁴ Aggressive management of hypertension in such patients can cause LVH to regress to the level seen in their nonhypertensive counterparts,¹⁵ thereby potentially improving outcome. Figure 29.2 demonstrates that there is a link between LVMI and eGFR even in the early stages of CKD.¹⁶

Atherosclerosis is a multisystem disease. CAD is the most common cause of HF, and 40% of patients with CKD have evidence of CAD.¹⁷ CKD is a proatherosclerotic condition and, in turn, atherosclerosis can lead to and exacerbate CKD. Vascular damage occurs at a microvascular level in the kidney and is very often independent of renal artery stenosis (RAS). The intrarenal arterial disease will lead to chronic glomerular damage. Smoking is independently associated with the development, and risk of progression, of CKD and the pathway is likely to involve the same atheromatous processes that contribute to CAD. The mechanisms by which CKD can exacerbate coronary artery atherosclerosis are discussed below.

Diabetic nephropathy is responsible for 20% of all new dialysis cases in the UK.¹⁸ Worldwide, this rises to as much as 40% in the United States and 55% in parts of the Indian subcontinent.¹⁹ In addition, up to 7.5% of the prevalent dialysis population have type 2 diabetes that developed after RRT was started.²⁰ Insulin resistance and chronic hyperglycaemia both lead to endothelial dysfunction, in which the usual antiatheromatous properties of vascular endothelium are disrupted. Subsequent macrovascular disease manifests most often as CAD and CKD. Microvascular disease is responsible for diabetic cardiac autonomic neuropathy and early endothelial disruption in the kidneys produces proteinuria which is a reliable marker of progressive CKD.²¹ Patients with a urine albumin-to-creatinine ratio of 30–299 mg/g, a range termed ‘microalbuminuria’, have a 9.2-fold risk of progression to established diabetic nephropathy compared to a control group of patients with no microalbuminuria.²²

Microalbuminuria is independently associated with a relative risk of cardiovascular mortality of 1.87 in diabetic patients (compared to those without albuminuria) and in a 10-year follow-up of diabetic patients with microalbuminuria, 9% of all-cause mortality was attributed to HF.²³ In the Heart Outcomes Prevention Evaluation (HOPE) trial, microalbuminuria was associated with an adjusted relative risk of 3.23 for hospitalization for HF. Importantly, this was similar for diabetic and nondiabetic patients, indicating a significant risk of HF for patients with other proteinuric illnesses (Table 29.4).²⁴

Diabetes mellitus is often part of the metabolic syndrome, in which there is coexistent obesity, hypertension, high triglycerides, and LDL cholesterol, and high circulating levels of prothrombotic/proinflammatory markers. These are all risk factors for cardiovascular and renal vascular damage, which emphasizes why diabetes is associated with such a high cardiorenal morbidity and mortality.

CKD can contribute to overactivation of the renin–angiotensin–aldosterone system (RAAS) which in turn both contributes to the development of cardiovascular disease and further exacerbates CKD, thereby initiating a pathway of progressive cardiorenal disease.

One of the primary purposes of the RAAS is to adapt to a drop in blood pressure in order to maintain vital organ blood flow. Angiotensin II causes sodium retention, expansion of the extracellular compartment, vasoconstriction, and restoration of organ perfusion. CKD causes chronic overactivation of the RAAS, and this appears to play a prominent pathophysiological role in the subsequent progression and exacerbation of intrarenal (and cardiac) damage.

Angiotensin II causes postglomerular arteriolar vasoconstriction, leading to intrarenal hypertension and glomerular damage. Its systemic vasoconstrictive activity also leads to LVH. Angiotensin II and aldosterone up-regulate activity of proinflammatory cytokines such as fibroblast growth factor (FGF), platelet-derived growth factor (PDGF) and transforming growth factor β₁ (TGFβ₁). These cytokines contribute to endothelial dysfunction, thereby promoting CAD and renal vascular disease, and are associated with intrarenal and myocyte fibrosis. Progression of CKD and HF can follow, ultimately resulting in an ever worsening cycle of progressive cardiorenal organ damage (Fig. 29.3).

The excess of vascular disease in patients with CKD is largely due to underlying risk factors. One particular problem relates to vascular calcification and associated arterial stiffness. There is a fivefold increase in vascular calcification of the coronary arteries in dialysis patients compared to other patients with coronary atheroma,²⁵ and there is evidence of more widespread ‘medial’ calcification within the arterial tree. This is again associated with an excess of mortality.²⁶ The calcification starts developing in the early stages of CKD and is present in over 50% of patients at the time of starting dialysis.²⁷ Once calcification is present, it continues to progress, though some medications have been shown to slow this progression. In some patients, renal transplantation will halt progression.²⁸

The pathophysiology of calcification is complex and involves an interplay between many predisposing factors including hyperphosphataemia, hypercalcaemia, and hyperparathyroidism, all of which can stimulate calcification of vascular smooth muscle cells and within the vascular matrix; CKD also leads to a reduction in endogenous inhibitors of calcification, such as fetuin A. Vascular calcification and renal bone abnormalities are now both encompassed by the term ‘chronic kidney disease–mineral bone disorder’ (CKD-MBD).

The clinical manifestation of calcification within the larger ‘conduit’ arteries is an increase in vascular stiffness which can be measured noninvasively with pulse wave velocity (PWV). An increase in PWV is associated with LVH and increased LVMI (and with reduced coronary filling), all of which may eventually predispose to HF.²⁹ Indeed, in some small studies, increased PWV, and thus vascular stiffness, has been shown to be more important than hypertension in the development of LVH.

That CKD-MBD represents a major cardiovascular risk highlights the importance of strict adherence to a ‘renal’ diet and appropriate use of phosphate binders at an early stage of CKD. However, there are possible risks of calcification associated with long-term use of high doses of oral calcium, as used in phosphate binders, although there is currently no proof of the association.

Endothelial dysfunction plays an important role in HF. It leads to a loss of the usual endothelium-mediated vasodilatory response to nitric oxide (NO) which is in turn an independent predictor of cardiovascular mortality in patients with HF. RAAS activation, mediated by CKD, exacerbates endothelial dysfunction by increased the production of reactive oxygen species (ROS).³⁰ Patients are less able to reduce afterload (which is increased by the vascular stiffness discussed above). Endothelial dysfunction has also been linked with abnormal myocardial remodelling and CAD.

CKD is a major contributor to endothelial dysfunction. Although an elevated serum creatinine, and its mathematical transformation into eGFR, is used to signify CKD, there is a host of other metabolites that are not routinely measured in CKD, many of which contribute to oxidative stress that in turn leads to end organ damage. Further, some circulating pro-inflammatory cytokines, such as interleukin-6 (IL-6), are normally excreted via the kidneys but cannot be removed by dialysis. These accumulate in advanced CKD and contribute further to the endothelial dysfunction.³¹ C-reactive protein (CRP) is a surrogate marker of the in vivo inflammation and is often raised in CKD, particularly in patients with tunnelled venous catheters as access for haemodialysis.³² Chronically elevated CRP in dialysis patients is associated with increased cardiovascular mortality, although not specifically from HF.³³

AKI as a primary event may lead to acute decompensated HF, even in patients with previously normal cardiac function. Retention of sodium and fluid occur as a result of renal injury and also as a result of attempts to resuscitate the unwell patient with intravenous fluid. When AKI supervenes, patients with underlying cardiovascular disease may also find that their medication, particularly diuretics and angiotensin converting enzyme (ACE) inhibitors, are stopped on admission to hospital, which increases the likelihood of a secondary episode of HF. Furthermore, electrolyte and metabolite disorders may develop which can further exacerbate decline in cardiac function. Specifically, renal acidosis is associated with pulmonary hypertension and right HF, and electrolyte disturbances leave patients at risk of arrhythmia and loss of effective atrial activity.

Patients with AKI and HF are difficult to manage. If oliguric AKI occurs and does not respond to early resuscitation, patients are at particular risk if acute pulmonary oedema or significant hyperkalaemia occur. They are unlikely to respond to diuretics, and so haemodialysis or haemofiltration are often necessary and should be considered early in the clinical course.

Chronic salt and water retention in advanced CKD can contribute to hypertension and ventricular dilation. In one study, Kayikcioglu et al.³⁴ managed hypertension in haemodialysis patients using dietary salt restriction and modification of dialysis target weight to control blood pressure, with no antihypertensive drugs. They achieved a target blood pressure of less than 140/90 mmHg in 90% of patients. Symptomatic hypotension was common but reduced to 7% of cases at 12 months. There was a reduction in LVMI from 164±64 to 112±36 g/m² during follow-up.³⁵

Anaemia is a risk factor for both the development of HF, and its poor outcome, independent of concurrent or causative CKD. Given the high incidence of anaemia associated with CKD, the contributory importance and treatment of anaemia has to be considered when managing the patient with CKD and HF. Figure 29.1 outlines the additive effect on mortality of coexistent anaemia, CKD, and HF.

An arteriovenous (AV) fistula is the preferred access for haemodialysis. It carries a three times lower risk of infection than tunnelled dialysis lines and there is a lower all-cause mortality in patients who dialyse via a fistula.³⁶ However, the creation of a shunt from the high-pressure arterial circulation into the lower-pressure venous system leads to circulatory changes and, ultimately, cardiac remodelling, which may cause high-output cardiac failure in a few individuals.

As little as 1 week after fistula formation, cardiac output may increase by up to 15%. There is an increased venous return and sympathetic activation with resultant resting tachycardia, and an increase in left ventricular end-diastolic volume, indicative of a greater circulating volume. The changes are thought in part to be a consequence of neurohormonal responses to the reduced vascular resistance that follows AV fistula formation, and patients with high-flow fistulae have higher circulating levels of natriuretic peptides because of the high volume state. There is also eccentric hypertrophy of the left ventricle in response to dilatation.³⁷ Although the majority of haemodialysis patients tolerate their fistula without any noticeable circulatory problems, a few patients are at risk of high-output cardiac failure. In such cases, ligation of the fistula must be considered as a therapeutic option. Male sex and use of proximal vessels for fistula formation are independent risk factors for the need for fistula ligation.

Atherosclerotic renovascular disease (ARVD) is common, and it is frequently seen in association with other cardiovascular diseases such as CAD, peripheral vascular disease, and stroke.³⁸ As would be expected, HF is also common in patients with ARVD. We found that ARVD was detectable in approximately one-third of elderly patients presenting acutely to hospital with HF.³⁹ In addition, 54% of a UK outpatient cardiac failure population had atherosclerotic renal artery stenosis (RAS) greater than 50%.⁴⁰ Conversely, HF is present in 38% of elderly US patients with ARVD⁴¹ and HF leads to an almost threefold increase in mortality risk compared to patients with ARVD but without HF.⁴² In many patients with ARVD and HF, there is normal LVEF, but patients with HF have higher filling pressures, higher LVMI, and greater prevalence of diastolic dysfunction than patients without HF.⁴² Systematic echocardiographic studies of cardiac structure and function in ARVD have shown that only 5% of patients had normal hearts, that the prevalence of LVH was twice as great as in CKD patients without ARVD, and that changes progress over time.^43,44

Patients admitted with decompensated HF have a worse outcome if there is associated AKI. A rise in serum creatinine of as little as 9 μmol/L is associated with both a prolonged inpatient stay and increased mortality.⁴⁵ The amount of change in creatinine is of more prognostic significance than the baseline creatinine.⁴⁶ In one study of 1004 patients admitted to hospital with decompensated HF, 25% had a rise in serum creatinine of more than 26.5 μmol/L (0.3 mg/dL). The presence of diabetes mellitus, hypertension, or CKD were independent predictors of postadmission AKI.⁴⁷ Anaemia, age, and the use of drugs blocking the RAAS and diuretics are predictors of AKI in the setting of acute hospitalization for HF. However, the precise pattern of risk is difficult to define because the definition of AKI varies from study to study. A 0.3 mg/dL rise in creatinine was defined as ‘worsening renal function’ (WRF) in the POSH study (the Prospective Outcomes Study in Heart failure). WRF was independently associated with higher serum creatinine levels on admission (odds ratio (OR) 3.02), pulmonary oedema (OR 3.35), but previous history of atrial fibrillation appeared to confer protection against WRF (OR 0.35). WRF was associated with an increase in average length of inpatient stay of 2 days but readmission rates and, importantly, mortality, were not affected.⁴⁸

The fact that AKI usually occurs very soon after hospital admission for HF suggests that low-output cardiac failure is important in its pathogenesis. Drugs blocking the RAAS lead to reduced renal perfusion and have a deleterious effect on renal function, due to actions on glomerular haemodynamics (that are beneficial when perfusion is better). However, hypotension and a low cardiac output state are not present in all patients who develop renal dysfunction. Other factors are important and there is an association between higher right atrial pressures and lower GFR.⁴⁹ Thus venous congestion and volume overload appear to contribute to AKI, a pathogenetic theory supported by experimental models in which temporary occlusion of renal veins leads to a temporary decline in GFR.⁴⁹ Similar pathological mechanisms may also account for the evolution of CKD in chronic HF given that lower ejection fraction does not correlate closely with the likelihood of progression of CKD, and CKD may occur in cases where there is normal LVEF.

The landmark trials of ACE inhibitors in patients with chronic HF excluded patients with significant renal impairment, and evidence for the use of blockers of the RAAS in patients with CKD and HF often comes from nonrandomized trials. For example, in a study of 1704 patients with systolic HF and CKD (mean eGFR 43 mL/min), ACE inhibitor was associated with an all-cause survival benefit of 4% compared to a propensity-score matched ‘control’ group not on ACE inhibitor. There was a similar reduction in all hospital admissions.⁵⁰ A rise in serum creatinine after starting an ACE inhibitor is not associated with a poor renal outcome, provided that the rise is not inexorable. In patients with HF, a rise in creatinine of up to 30% followed by stable renal function should be accepted and should not lead to stopping the drug or further investigation, whereas a greater rise raises the possibility of ARVD: renal imaging may be indicated in selected cases.

The benefit of ACE inhibitors in patients with ESRD but no HF is less clear. A theoretical benefit is the potential for reduction in left ventricular mass given that LVH is associated with the risk of progression to cardiac dilation and HF, and of course, mortality.³ However, ACE inhibitors have been shown to improve survival as a secondary endpoint in observational studies of dialysis patients.⁵¹ As ACE inhibitors may cause hyperkalaemia, electrolyte changes should be very closely monitored in dialysis patients

Angiotensin II receptor blockers (ARBs) have a similar beneficial profile in CKD to ACE inhibitors. They also have a similar profile of adverse effects, so that converting from ACE inhibitor to ARB in cases of hyperkalaemia is unlikely to improve matters; however, a change may be indicated in cases of cough related to the use of an ACE inhibitor. Aliskiren, the first commercially available drug to inhibit renin directly, can improve neurohumoral markers of HF,⁵² but its effects on clinical outcomes have not been fully evaluated.

Spironolactone, an aldosterone receptor antagonist, improves survival in HF⁵³ but at the expense of increasing risk of hyperkalaemia, and its use in patients with CKD is thus limited. The RALES trial excluded patients with significant renal impairment.⁵³ Dual blockade of the RAAS using a combination of two classes of drugs is commonplace in nephrology and is generally considered safe for treatment of HF.⁵⁴ An important exception is dual blockade including spironolactone in patients with CKD stages 4 and 5, in whom the risk of hyperkalaemia is considerable.

β-Blockade confers significant survival benefit in patients with stage 3 CKD, CAD, and HF (OR 0.75 vs matched patients not on β-blockade).⁵⁵ Dialysis patients with HF also benefit, (2-year all-cause mortality, carvedilol vs placebo 51.7% vs 73.2%, p 〈 0.001)⁵⁶ and are less likely to develop de novo HF if pre-emptively prescribed a β-blocker (OR 0.69 vs matched patients not on β-blocker).⁵⁷ However, β-blockers appear to be underused in patients with CKD compared to matched patients without renal impairment,⁵⁵ despite poor renal function having little adverse effect on the efficacy of β-blockade in treatment of HF. In elderly patients with a low GFR there is a modest tendency to bradyarrhythmias leading to drug discontinuation (2.3% vs 0.8% in placebo).⁵⁸

As is the case with β-blockers, the efficacy of digoxin is not affected by CKD and it does not correlate with GFR.⁵⁹ However, digoxin toxicity is more common in patients with CKD as the drug is partially eliminated by the kidney; for those patients with ESRD, the drug is not removed by dialysis. Digoxin can also contribute to hyperkalaemia and drugs such as calcium gluconate, used in the management of hyperkalaemia, can exacerbate the arrhythmic risk of digoxin toxicity. As digoxin does not confer the same survival benefit as β-blockers in HF, it should not be considered a first-line therapy for HF in patients with CKD.

The principal of removing ascites to treat HF was first demonstrated 60 years ago. The possibility of using peritoneal dialysis as a management strategy for chronic HF in patients with and without CKD is now also gaining support. In one small study (n = 17, mean age 64±9 years), patients with refractory HF who were started on peritoneal dialysis had a significant survival benefit (82% 12-month survival) and fewer hospital admissions (reduction from 62±16 to 11±5 days per patient per year).⁶⁰ In refractory HF, the ability of the kidneys to generate a diuresis is blunted, as is their response to large-dose diuretics. Offloading fluid from the circulation by removal during peritoneal dialysis will improve haemodynamics—for example, by reducing right atrial pressure—so improving cardiac function and renal perfusion. There is a trend towards offering peritoneal dialysis rather than haemodialysis to patients with CKD stage 5 and coexistent HF. The theory is that patients with markedly impaired ventricular function will benefit from gradual fluid removal and so be less likely to suffer collapse relating to hypotension, and/or serious arrhythmias that might result from haemodialysis. However, the evidence to support this practice is conflicting and derived from small, nonrandomized studies.

One drawback that may prevent more widespread application of peritoneal dialysis (PD) in patients with HF and earlier stages of CKD is that patients need to be mobile and quite physically able to manage the technique; even where assisted PD programmes are available, the patient will be expected to manage all or some of their therapy. A 5-L bag of dialysis fluid weighs 5 kg and carrying two or more of these to a PD machine is an effort for any dialysis patient.

Treatment of anaemia with erythropoiesis-stimulating agents (ESAs) and intravenous iron is associated with an improvement in cardiac function, and symptom and diuretic burden in HF. The effect is less marked in such patients who also have CKD.⁶¹ The symptom benefit may not equate to survival benefit. TREAT (Trial to Reduce Cardiovascular Endpoints with Aranesp Therapy) aimed to show a benefit in using ESAs to achieve a higher than usual target haemoglobin (13 g/dL) in diabetic predialysis CKD patients. However, there was an excess of thromboembolic events and cerebrovascular events in the ESA arm compared to placebo.⁶² The FAIR-HF trial⁶³ compared intravenous iron therapy with placebo in patients with NYHA class II or III HF and biochemical evidence of iron deficiency, but excluded patients with renal impairment (mean eGFR 64 mL/min). It showed a functional and symptomatic benefit without the adverse event profile of TREAT—there was a small but statistically significant improvement in haemoglobin in the treatment arm (13.0 ± 1 g/dL vs 12.5 ± 1 g/dL in placebo, p 〈 0.001).

The nephrotoxicity and decline in renal function with diuretic therapy may cause a therapeutic dilemma. Patients with decompensated HF with either associated AKI or underlying CKD are more likely to suffer further AKI due to diuretic drugs, and the adverse effect is even more pronounced in patients receiving blockers of the RAAS. ACE inhibitors (or ARBs) may have to be stopped during an acute admission due to AKI or hyperkalaemia, but they should be reintroduced when the acute risk has passed.

Patients with AKI or CKD are less responsive to diuretics than patients without renal impairment. The use of high doses of diuretics is correlated with poor outcome in HF, and the association is most likely due to the diuretic resistance seen in patients with the more severe forms of the cardiorenal syndrome, and in those with very poor haemodynamic status, in turn a marker of the most severe HF. It is important to note that using diuretics to improve urine output in acutely unwell patients will not improve renal function (in fact it may worsen it in those with hypovolaemia), and so diuretics should only be used when the removal of excesses of fluid is clinically required. Therapeutic B-type natriuretic peptide (BNP, nesiritide) does not reduce the need for diuretic, nor increase urine output in patients admitted acutely with HF.⁶⁴ Nitrates, and also hydralazine, are useful alternatives to blockade of the RAAS when HF exacerbations are complicated by AKI. However, hypotension may preclude their use.

In the acute setting, if nitrates are unable to stabilize a patient with severe HF coupled with AKI and diuretic resistance, the treatment options become quite limited, at which point haemofiltration may need to be considered (see below).

Positive inotropic agents were once regularly used for treating acute admissions with decompensated HF, but this practice has now fallen from favour. Low doses of dopamine interact with specific receptors in the kidney resulting in an increase in renal blood flow without apparent significant inotropic effect, and hence dopamine continues to be used by some at a ‘renal dose’, with the aim of improving the response to diuretics, and reducing the incidence and progression of AKI.⁶⁵ Unfortunately any data which has shown benefits of low dose dopamine regimes have derived from small-scale selected nonrandomized studies.

Haemofiltration is the ultrafiltration of fluid from the body via an extracorporeal machine. It uses the same principle as that used to offload fluid from haemodialysis patients, and it can now be achieved with portable machines designed specifically for haemofiltration only. Ultrafiltration has the advantage over diuretic therapy in that more rapid fluid and sodium removal is possible during the early phase of an admission with decompensated HF, minimizing the tendency to deterioration of renal function or hypotension that might accompany high dose diuretic therapy. The controlled nature of the haemofiltration process may thus benefit patients with labile blood pressure response during treatment. Studies have shown a short-term improvement in outcome when using haemofiltration instead of intravenous diuresis as first-line therapy for acute admissions with HF,⁶⁶ but longer-term outcome results are still awaited.

There is little evidence to support the use of PD in the management of acute HF. The logistics and safety of inserting a PD catheter in an acutely unwell patient, followed by its immediate use, are likely to preclude its becoming a standard therapy.

The evidence behind the use of BNP as a diagnostic tool has come from studies that either excluded patients with significant CKD, or studies that showed reduced clinical efficacy of BNP in the context of coexistent CKD. Given the prevalence of kidney disease in the HF population and vice versa, it is vital to have better understanding of BNP in such scenarios. Ventricular hypertrophy, dilatation, and abnormalities of function, as well as extracellular fluid accumulation, are commonplace in CKD patients, making the diagnosis of HF difficult. BNP is a potentially useful clinical tool in these circumstances.

Difficulties in interpretation arise because BNP levels may be elevated in CKD in the absence of HF. BNP correlates strongly with LVH, commonplace in CKD. BNP is partly metabolized in the renal parenchyma, so levels will also tend to be elevated in patients with reduced renal function. It is not removed by dialysis, but posthaemodialysis BNP levels show a fall that correlates with ultrafiltration volume. An elevated BNP is less likely to occur in patients using PD. BNP is probably still useful in diagnosing HF in ESRD, but only if higher diagnostic cut-offs (e.g. 200 pg/mL) are used, which reduces the sensitivity of the test. Its use should not replace clinical judgment.^64,67,68

Although BNP increases renal perfusion, its therapeutic use does not reduce the occurrence of AKI in decompensated HF in patients with pre-existent CKD.⁶⁹ BNP reduces distal tubular reabsorption of sodium, leading to effective diuresis, an effect which is lost in CKD where glomerular or tubular disease predominate. The blunting of this effect may contribute to the high occurrence of HF in CKD patients, as such patients are less able to mount an effective response to fluid overload.

A small number of patients with bilateral significant renal artery stenosis (RAS) present with sudden onset ‘flash’ pulmonary oedema, which can be life threatening.⁷⁰ These patients can be recognized by the absence of overt CAD, and the presence of severe, often accelerated, hypertension, AKI, and evidence of widespread atheroma (particularly arterial bruits). The Angioplasty and Stenting for Renal Artery Lesions (ASTRAL) trial has shown that endovascular intervention (angioplasty and stenting) with medical therapy does not reduce cardiovascular events and mortality in over 800 patients with significant RAS, when compared to medical therapy alone.⁷¹ However, the study did not examine the effect of revascularization on HF, and nor has any previous trial. Despite the absence of conclusive benefit, the clinical consensus is that patients with ‘flash’ pulmonary oedema and RAS should be urgently treated with endovascular renal revascularization and there are many reports of subsequent successful clinical outcomes.⁷²

Heart failure in CKD carries an excess mortality through a heterogeneous series of pathophysiological interactions. Prevention of LVH through blood pressure management and volume control, and early modification of risk factors for CAD is vital. A change in eGFR and proteinuria are early indicators of adverse cardiovascular outcome and, thus also indicate the need for early therapeutic intervention. Simple medical therapy, such as β-blockers and drugs that act on the RAAS, are effective but underused in patients with HF and renal impairment. These agents are safe provided that patients are appropriately monitored. Management of HF in the setting of acute or chronic kidney disease is complicated by the potentially nephrotoxic effect of many current therapies. PD and haemofiltration are measures that may yet improve the outcome of cardiorenal disease.