Heart failure with preserved ejection fraction (HFpEF) is increasingly prevalent across the world and comprises around half of all patients with clinical heart failure [1]. Diagnosis of HFpEF is considered in patients who are breathless, do not have reduced left ventricular ejection fraction (LVEF) (commonly taken as $\ge$50%, but cut-off varies), and in whom other cardiac and non-cardiac causes for symptoms have been excluded, and consensus diagnosis recommendations have now been developed [2]. Chronic kidney disease (CKD) encompasses structural kidney abnormalities or reduced function over at least three months, and is also a growing global burden with a global prevalence of 9.1% [3]. Both HFpEF and CKD are increasingly common at older ages, share a number of risk factors for development, and dual diagnosis is common and set to grow. In this paper we will review the pathophysiological links between HFpEF and CKD, prevalence and prognosis of dual diagnosis, evidence for pharmacological treatment options, and practical considerations for management of the co-existing conditions, and briefly review acute kidney injury (AKI) in patients with HFpEF.

While common risk factors for HFpEF and CKD are clearly identified, including diabetes, hypertension and obesity, a number of pathophysiological processes contribute to the interplay between cardiac and renal dysfunction (Fig. 1), which are not completely understood.

Fig. 1.

Summary of pathophysiological processes linking heart failure with preserved ejection fraction (HFpEF) and chronic kidney disease (CKD). HFpEF and CKD share common risk factors including diabetes, hypertension and obesity, but also a number of pathophysiological processes contribute to the interplay between cardiac and renal dysfunction.

Myocardial fibrosis is the key pathology seen in HFpEF [4], with resulting cardiac structural and functional abnormalities which can potentially affect kidney function. Right ventricular (RV) dysfunction has been implicated in the development of renal impairment in patients with heart failure with a reduced ejection fraction (HFrEF) for some time [5]. In HFpEF worse RV indices are seen in association with CKD, often with poorer prognoses [6, 7, 8]. Left ventricular (LV) hypertrophy without dilatation is a further cardinal feature of HFpEF, and increased LV mass predicts progression of renal disease in community cohorts at high risk for heart failure [9, 10]. Left atrial enlargement and diastolic dysfunction are common in HFpEF [11, 12, 13], and impaired left atrial function predicts renal function decline in patients seen in a tertiary cardiology clinic with dyspnoea [14] and left atrial strain predicts progression of renal disease in patients with CKD [15], although causality is uncertain.

Pulmonary hypertension is seen in the majority of patients with HFpEF [13], largely thanks to metabolic insult in addition to pressure-related injury [16]. Haemodynamic markers of pulmonary hypertension (elevated transpulmonary gradient, pulmonary vascular resistance, and diastolic pulmonary gradient) were associated with increased mortality and increased cardiac hospitalisations in patients with HFpEF from a large cohort undergoing right-heart catheterisation [17]. Pulmonary hypertension is recognised as increasing the risk of poorer outcomes after kidney transplantation [9, 10, 18, 19].

Furthermore, increased central venous pressure in HFpEF results in reduced glomerular capillary blood flow and raised interstitial and tubular pressures within the kidney [20]. On top of that, patients with HFpEF commonly have chronotropic incompetence, which is independently associated with reduced glomerular filtration rate (GFR) [21], likely due to renal hypoperfusion. Novel haemodynamic parameters may offer insights into further pathophysiologic mechanisms [22].

In the opposite direction, renal impairment is an independent risk factor for development of HFpEF [23]. Indeed, in a porcine model CKD via 5/6 nephrectomy induced development of HFpEF [24], while human data suggest that advanced CKD is associated with a variety of cardiac structural abnormalities [25]. Patients with CKD have impaired sodium handling and fluid overload, directly contributing to venous congestion, which is more pronounced in the advanced stages of the disease. Up to date understanding makes central the importance of systemic inflammation in HFpEF pathophysiology, leading to coronary microvascular endothelial inflammation and dysfunction, and ultimately to stiff cardiomyocytes and diastolic LV stiffness via reduced nitric oxide signalling [26]. Traditional HFpEF co-morbidities such as diabetes and hypertension contribute to the inflammation, and CKD itself is well-recognised as a pro-inflammatory state [27]. Obesity is over-represented in older patients with HFpEF and adipose tissue contributes to a number of pathophysiological pathways increasing inflammation and oxidative stress, potentially leading to both cardiac and renal dysfunction [28]. Further, there is some evidence that increased fibroblast growth factor 23 levels in CKD induce LV hypertrophy [29], while enhanced sympathetic activation in CKD can contribute to heart failure [30], with increased neurohormonal activation noted in both CKD and pulmonary hypertension being associated with vascular remodelling and worse outcomes [31].

It is also worth considering that in a minority of patients HFpEF may be caused by specific diseases, some of which, such as amyloidosis, may affect the kidneys concurrently.

The European Society of Cardiology 2019 consensus statement on the diagnostic criteria for HFpEF includes the definition of a “preserved” LVEF as $\ge$50% [2]. HFpEF was historically described as “diastolic heart failure”, and the threshold for EF has not always been consistent. It is necessary when considering HFpEF studies and treatment trials to acknowledge these inconsistencies and highlight the HFpEF inclusion criteria and relevant EF-based sub-group analysis. We include a table of the most important trials discussed below, with the key details for each (Table. 1).

CKD is commonly seen in HFpEF, although estimated prevalence varies. Most studies report CKD as estimated GFR (eGFR) 60 mL/min/1.73 m${}^2$, corresponding to stage 3 CKD onward [32]. Recent studies have reported rates of CKD in HFpEF as 60% in a Belgian cohort [33], 50% in a pan-Asia cohort [34] (both including LVEF $\ge$50%), and 48% in the worldwide PARAGON-HF study [35] (LVEF $\ge$45%). Albuminuria is also common in heart failure even without reduction in eGFR [36], making it likely that CKD is generally under-reported.

Multiple studies have shown that HFpEF patients with CKD have higher rates of traditional cardiovascular risk factors such as hypertension and diabetes than those without [6, 7, 33, 37], and the clinical picture of patients with HFpEF and co-existing CKD is generally that of more advanced disease.

HFpEF encompasses a broad range of clinical characteristics, possibly resulting from the multiple pathophysiological mechanisms, making understanding and prediction of renal involvement somewhat challenging. Phenomapping has allowed for distinct clusters of HFpEF patients to be classified. The first published study found that the cluster more likely to have CKD are older and have most severe electrical and myocardial remodelling, pulmonary hypertension, and RV dysfunction [8].

The cohort with CKD (eGFR 20–60 mL/min/1.73 m${}^2$) in the RELAX study (inclusion LEVF $\ge$50%) had more advanced symptoms, more impaired indices of cardiac function and structure, and higher levels of biomarkers of neurohormonal activation, collagen metabolism/fibrosis and oxidative stress, when compared to those with normal kidney function. They also had lower exercise capacity, which was largely accounted for by lower haemoglobin levels [7].

A prospective study of 299 HFpEF patients (with LVEF $>$50%) in the USA used conventional 2-D and speckle-tracking echocardiography to characterise cardiac structural and functional indices in HFpEF patients with and without CKD. As seen in other cohorts those with CKD (taken as 60 mL/min/1.73 m${}^2$) were older, had a higher beta natriuretic peptide (BNP), were more likely to have diabetes and hypertension, and had a worse New York Heart Association (NYHA) functional class. Parameters of LV longitudinal strain, RV free wall strain and left atrial (LA) strain were associated with CKD, even after multivariate adjustment [6].

HFpEF is also common in patients with end-stage renal disease (ESRD) managed with maintenance haemodialysis (HD) or peritoneal dialysis (PD). Antlanger et al. [38] reviewed a cohort of 105 ESRD patients undergoing regular HD in Vienna, Austria, and found that 57% fulfilled the definition criteria of HFpEF (including LVEF $\ge$45%), compared to 13 with HFrEF (LVEF 45%). Predictors of HFpEF in this cohort included higher age, female sex, higher systolic blood pressure, and dialysis vintage, as we might expect, but interestingly not diabetes. Fluid overload (assessed by pre-dialysis bioimpedance spectroscopy) was not associated with heart failure, and nor were dialysis-related factors including access via arteriovenous fistula, blood flow, and ultrafiltration volume. Of a further cohort of 1034 regular HD patients in Romania, 266 (26%) had HFpEF (LVEF $\ge$50%), compared to 118 with heart failure with mid-range ejection fraction (HFmrEF, LVEF 40–49%) and 38 (4%) with HFrEF (LVEF 40%), although characteristics of the specific groups are not given [39]. Even without overt heart failure, structural heart abnormalities are common in ESRD. In a study of 654 patients starting HD, 567 (87%) had echocardiographic evidence of structural heart disease, with diastolic dysfunction in 350 (78% of the 448 in which diastolic function was reviewed) [25].

In a Chinese cohort of ‘advanced CKD’ patients – in practice a mix of CKD stage 4 and ESRD, including those on HD and PD – the HFpEF group (inclusion LVEF $\ge$50%) were more likely than the non-HFpEF group to have diabetes and also had a higher systolic blood pressure [40]. Finally, among 220 patients undergoing regular PD in Hong Kong, 47 (21%) had HFpEF (inclusion LVEF $\ge$50%) and 39 (18%) had HFrEF [41]. In this cohort higher age, presence of diabetes, background of coronary artery disease, and/or other atherosclerotic vascular disease, higher BNP and higher troponin T were all associated with HFrEF, with the HFpEF group generally intermediate between HFrEF and no heart failure. There were no differences observed in PD parameters.

Overall, patients with concomitant HFpEF and CKD appear to be characterised by a more severe disease burden as a result of both the presence of shared for the two conditions risk factors as well as the added impact of the two conditions.

It has been recognised for some time that prognosis is worse for HFpEF patients with concomitant renal dysfunction. Studies of patients admitted with acute decompensation of HFpEF report rates of baseline CKD far higher than those seen in the general HFpEF population, of up to 82% [42].

In a cohort of 397 patients identified post-discharge following hospital admission with symptomatic HFpEF, the pheno-group with higher rates of CKD had increased mortality and hospitalisation rate than those without [8]. The diagnostic cut-off for LVEF was $\ge$50%, the mean reported LVEF was 61%, without significant differences between pheno-groups. In a Belgian cohort of 183 patients HFpEF (inclusion LVEF $\ge$50%) and multivariate analysis showed that lower eGFR independently predicted the composite endpoint of all-cause mortality or hospitalisation for heart failure, and mortality alone [33]; this cohort was notably elderly (mean age 78 $\pm$ 8 years) and excluded those with NYHA functional class I disease.

The Angiotensin Receptor Neprilysin Inhibition in Heart Failure With Preserved Ejection Fraction (PARAGON-HF) trial compared treatment with sacubitril/valsartan versus valsartan in 4796 patients with HFpEF over a median follow-up period of 34 months. Participants were all aged $\ge$50 years and had NYHA functional class II-IV symptoms. Analysis of those enrolled showed a higher all-cause mortality risk for patients with eGFR 60 mL/min/1.73 m${}^2$ compared to with eGFR $\ge$60 mL/min/1.73 m${}^2$ [43]; it is important to note that patients with eGFR 30 mL/min/1.73 m${}^2$ were excluded from the trial. PARAGON-HF also defined HFpEF as including participants with LVEF $\ge$45%, although mean LVEF of enrolled participants was 57.6%, and based on the presented standard deviation in each arm only a very small % of participants would have had LVEF less than the more usual 50% [35].

In line with PARAGON-HF findings, the analysis of the 1767 patients in the Americas group in the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist Trial (TOPCAT) study highlighted an independent inverse relationship between eGFR and many clinical outcomes, including all-cause mortality and heart failure hospitalisations [37]. Again, patients with an eGFR 30 mL/min/1.73 m${}^2$ were excluded, and HFpEF threshold included patients with LVEF $\ge$45%. A recent secondary analysis of the TOPCAT cohort also observed that longitudinal variability of kidney function laboratory parameters (urea and creatinine) was independently associated with mortality and the composite endpoint of cardiovascular death, heart failure hospitalisation and aborted cardiac arrest [44]. Chirinos et al. [45] analysed samples from TOPCAT to consider potential prognostic biomarkers. For consideration of kidney function or injury, the panel include Neutrophil gelatinase-associated lipocalin (NGAL), Cystatin-C, and Kim-1, but none of these were predictive of the composite endpoint.

Proteinuria or albuminuria are signs of abnormal kidney function and can often be present before GFR drops. As such, a proportion of patients with CKD may be overlooked if the only criteria used to identify them is eGFR $\le$60 mL/min/1.73 m${}^2$. Accordingly, some studies have specifically assessed their impact. Albuminuria without reduced GFR was associated with poorer outcomes and cardiac remodelling in a US cohort of 144 patients with HFpEF including LVEF $\ge$50% [46]. Further, the presence of dipstick proteinuria and albuminuria appear to be independent prognostic markers in HFpEF [47, 48].

Information regarding prognosis in patients with concomitant ESRD and HFpEF is limited. In one small study of patients on regular HD, the presence of HFrEF was strongly associated with a higher risk of cardiac death and/or hospitalisation for cardiovascular reason adjusted hazard ratio (HR) 3.24; 95% CI 1.08–9.75; p = 0.037) but for HFpEF (including LVEF $\ge$45%) this association did not reach statistical significance (adjusted HR 4.31, 95% CI 0.46–40.03; p = 0.199) [38]. A recent study of 1034 patients on regular HD in Romania found comparable survival (all-cause mortality) to four years between the groups with HFpEF and no heart failure, but reduced survival for patients with LVEF 50% [39]. In PD, analysis of outcomes over four years showed that HFpEF (including LVEF $\ge$50%) was associated with increased cardiac mortality, all-cause mortality, heart-failure events and cardiovascular events compared to those without heart failure, but for all outcomes the risk was significantly greater for those with HFrEF [41].

Overall, renal impairment appears to be associated with a worse prognosis in HFpEF patients. There is not specific data regarding the prognostic implications of HFpEF on CKD progression or mortality; nevertheless, a number of echocardiographic features typically seen in HFpEF have been associated with poorer renal outcomes and/or mortality in CKD [49].

Acute Kidney Injury (AKI) is defined by a sudden deterioration in renal function with rapid increase in serum creatinine or decrease in urine output [90], and is common in patients admitted to hospital for treatment of acute decompensation of heart failure (ADHF). A large registry study in the USA found that AKI complicated 20% of two million admissions with ADHF patients without underlying CKD [91], and slightly more commonly seen in HFrEF than in HFpEF.

In a pooled analysis of four studies of HFpEF patients admitted to hospital with ADHF, baseline renal dysfunction, defined as eGFR 60 mL/min/1.73 m${}^2$) was present in 82% of participants [92]. Calculating from the figures presented, AKI occurred in 22% of admissions. Obese patients admitted with ADHF were younger, had higher blood pressure and were more likely to have diabetes, but had similar baseline kidney function to non-obese counterparts. During comparable diuresis obese patients were significantly more likely to suffer AKI.

It is difficult to be sure whether transient creatinine changes during diuresis reflect changing haemodynamics within the glomerulus or true organ damage. An analysis of undifferentiated (i.e., both HFrEF and HFpEF) patients undergoing aggressive diuresis during ADHF did not find associations between worsening renal function and recognised markers of renal tubular injury including neutrophil gelatinase-associated lipocalin (NGAL) and kidney injury molecule 1 (KIM-1) [93]. Worsening renal function during ADHF is associated with echocardiographic signs of increased right ventricular free wall thickness and right ventricular dysfunction [42].

AKI in general is associated with future development of CKD [94], although long-term prognosis following AKI in HFpEF is not yet demonstrated. While in an undifferentiated cohort admitted with ADHF worsening renal function was associated with significantly worse in hospital mortality, complication rate and length of stay [95], in a pure HFpEF cohort of 332 patients (including with and without CKD) 30 day outcomes were not affected by development of AKI [92]. Patients with HFpEF with normal baseline renal function admitted with ADHF who developed AKI had an increased in-hospital mortality: 4.9% vs. 1.6%, adjusted odds ratio (OR) 3.21, p $\le$ 0.001 [91].

Cardiac catheterization, commonly used for diagnostic and therapeutic reasons, carries a risk of post-procedure AKI. While there are clear recommendations for pre-procedure hydration to mitigate AKI risk in patients without heart failure, and at a lower rate for those with HFrEF, the little evidence available to date suggests that over-hydration is not helpful for those with HFpEF [96].

HFpEF and kidney dysfunction commonly co-exist, and cardiologists, nephrologists and community physicians will care for patients with both with increasing frequency. They commonly share multiple risk factors and specific pathophysiological features of each contribute to the development and progression of the other, to the extent that they may often be seen as different manifestations of the same disease spectrum. Patients with a dual diagnosis of HFpEF and CKD appear to bear a more severe clinical picture and a worse prognosis. Neither CKD as caused by diabetes/hypertension/obesity nor HFpEF currently have proven treatments beyond control of co-morbidities and management of symptoms, and prevention and limitation of progression are key. The emerging evidence of a potential promising role of SGLT2 inhibitors in the management of HFpEF, on the background of their already demonstrated renal prognostic benefit may significantly add to the limited existing strategies, and possibly provide valuable insights on disease pathophysiology, should targeted research be taken forward.

JJ wrote the draft. EL conceived the idea and edited. EA edited the final version of the manuscript. All authors read and approved the final version of the manuscript.

Not applicable.

Not applicable.

This study received no external funding.

The authors declare no conflict of interest. Eirini Lioudaki and Emmanuel Androulakis are serving as the Guest editor of this journal. We declare that Eirini Lioudaki had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to Giuseppe Coppolino.