Cardiorenal protection in advanced chronic kidney disease: research highlights from landmark papers published in Nephrology Dialysis Transplantation during 2018

Chronic kidney disease (CKD) and cardiovascular disease go hand-in-hand and an inverse, strong and independent relationship exists between estimated glomerular filtration rate (eGFR) and the risk of cardiovascular morbidity and mortality [1, 2]. In fact, the mortality hazard attributable to CKD is even higher than the mortality risk arising from diabetes mellitus (DM) and comparable with that of established coronary artery disease, thus CKD is considered a coronary disease equivalent [3]. The burden of cardiovascular disease, particularly due to heart failure or the development of cardiac arrhythmias, is multiplied in those with end-stage renal disease (ESRD) on chronic dialysis. It has to be noted, however, that clinical trial evidence to guide the management of cardiovascular disease in these high-risk patients is scant [4]. This lack of evidence is at least partly attributable to the fact that advanced CKD/ESRD was systematically an exclusion criterion in the majority of large outcome trials that have demonstrated the safety and efficacy of several cardioprotective therapies in earlier stages of CKD or in the general population [4].

In this editorial we discuss therapeutic strategies to optimize cardiovascular and/or renal protection in advanced CKD based on research highlights from ‘landmark’ papers published in Nephrology Dialysis Transplantation during 2018. We provide ‘take-home messages’ from these studies and directions for future research in these important areas.

RENIN–ANGIOTENSIN–ALDOSTERONE SYSTEM (RAAS) BLOCKADE IN PATIENTS WITH DM AND STAGES 3b–5 CKD

Based on the evidence from randomized trials showing that among patients with DM and overt nephropathy, agents blocking the RAAS are superior to other antihypertensive drug classes in retarding the progression of kidney injury [5, 6], guidelines recommend the use of RAAS blockers as first-line antihypertensive therapy in diabetic kidney disease [7, 8]. Nistor et al. [9] challenged the scientific basis of this guidance, particularly for patients with DM and advanced (Stages 3b–5) CKD. They reasoned that in this particular high-risk subgroup, the potential cardioprotective action of RAAS blockers, if any, might be counteracted by excess risk of serious renal adverse events (i.e. acceleration of eGFR decline, hyperkalaemic events and premature initiation of dialysis). To explore this provocative research hypothesis, Nistor et al. [9] conducted a meta-analysis of randomized trials with a follow-up of at least 6 months comparing monotherapy with a RAAS blocker versus placebo or active treatment with other antihypertensive drug classes. Randomized trials were eligible in this meta-analysis if they reported treatment effects in patients with DM and moderate-to-severe CKD (defined as baseline eGFR <45 mL/min/1.73 m²). This meta-analysis also included post hoc analyses of randomized trials if their statistical methodology maintained the intention to treat principle or if there was a report of an interaction analysis for participants with a baseline eGFR <45 mL/min/1.73 m² [9].

Seven randomized trials and four post hoc analyses of randomized trials (incorporating data from 9979 participants) fulfilled the above-mentioned inclusion and exclusion criteria. Contrary to the original hypothesis, this meta-analysis showed that compared with placebo or other antihypertensive agents, RAAS blockade lowered by 19% the composite outcome of doubling of serum creatinine or the need for dialysis {relative risk [RR] 0.81 [95% confidence interval (CI) 0.70–0.92]} [9]. This improvement in the composite renal outcome was accompanied by a 10% reduction in the risk of non-fatal cardiovascular events [RR 0.90 (95% CI 0.81–1.00)] [9]. Since heart failure is the major cardiovascular morbidity in patients with advanced CKD, this benefit would probably be even greater if hospitalizations for heart failure had been taken into consideration in the assessment of the composite outcome of non-fatal cardiovascular events. What is good for the kidney is also good for the heart—in other words, a therapy that protects the kidney is likely to also be protective against heart failure hospitalization.

In fact, the above meta-analytic evidence is not opposing but is supportive of the guidance to use RAAS blockers as first-line antihypertensive therapy in diabetic kidney disease. It has to be noted, however, that the opportunity to follow the guideline recommendations and offer optimal cardiorenal protection in the population of interest in this meta-analysis (i.e. in those with DM and eGFR <45 mL/min/1.73 m²) is often missed, owing to the risk of inducing hyperkalaemia with RAAS blockade [10]. Unfortunately, the incidence of hyperkalaemia or the discontinuation rate of RAAS blockers due to overt hyperkalaemia was not quantified in this meta-analysis because of incomplete reporting of these adverse events in studies included in the quantitative data synthesis [9]. The hyperkalaemia risk associated with RAAS blockade use, however, is high [10] and mandates the careful monitoring of serum potassium levels after initiation of RAAS-blocking agents, particularly in such high-risk patients.

Newer therapies that bind potassium in the gut and enhance its removal may overcome the barrier of hyperkalaemia and enable the wider use of RAAS blockade for cardiorenal protection in advanced CKD [10]. Some proof of this concept is provided by the recently published AMBER trial [11]. In this trial, 295 patients with uncontrolled resistant hypertension and advanced CKD (eGFR 25–45 mL/min/1.73 m²) were randomized to double-blind therapy with the potassium-binder patiromer (8.4 g/day) or placebo, in addition to open-label spironolactone (at a starting dose of 25 mg/day) and their background antihypertensive medications. Over a 12-week follow-up, the proportion of participants who continued to receive spironolactone was higher in patiromer-treated than in placebo-treated participants [between-group difference 19.5% (95% CI 10.0–29.0)] [11]. Thus just five patients treated with patiromer would enable the use of spironolactone in one additional patient.

SODIUM–GLUCOSE CO-TRANSPORTER 2 (SGLT-2) INHIBITION IN ADVANCED DIABETIC KIDNEY DISEASE

The Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation (CREDENCE) trial randomized 4401 patients with type 2 DM and overt nephropathy (eGFR 30–90 mL/min/1.73 m² and albuminuria >300 mg/g) who were already receiving guideline-based therapy with a RAAS blocker to receive double-blind canagliflozin (100 mg/day) or placebo [12]. At a median follow-up of 2.6 years, CREDENCE was prematurely terminated due to an impressive 30% reduction in the occurrence of the primary composite outcome of ESRD, doubling of serum creatinine or death from renal or cardiovascular causes with the SGLT-2 inhibitor canagliflozin [12]. Compared with placebo, canagliflozin also provoked a 20% reduction in the risk of cardiovascular death, myocardial infarction or stroke as well as a 39% reduction in the risk of hospitalization for heart failure [12]. The publication of these impressive results was followed by a reappraisal of the 2019 American Diabetes Association guidelines [13]. SGLT-2 inhibitors are now recommended for patients with diabetic kidney disease and an eGFR ≥30 mL/min/1.73 m², particularly in the presence of albuminuria >300 mg/g [13]. The indication to use SGLT-2 inhibition is to offer optimal cardiorenal protection and not to improve glycaemic control with these agents. It has to be noted, however, that since patients with DM and eGFR <30 mL/min/1.73 m² were excluded from the CREDENCE trial, the safety and efficacy of SGLT-2 inhibitors in more advanced diabetic kidney disease are not yet clearly defined.

In a combined analysis of 11 Phase 3 randomized trials (incorporating data from 220 patients with type 2 DM and eGFR ranging from 12 to 45 mL/min/1.73 m²), Dekkers et al. [14] compared the albuminuria-lowering effect of the SGLT-2 inhibitor dapagliflozin, administered at doses of 5 and 10 mg/day, versus placebo in Stages 3b–4 diabetic kidney disease. The mean overall eGFR at baseline was ∼38.0 mL/min/1.73 m² (∼20 mL/min/1.73 m² lower than the baseline eGFR of CREDENCE participants). Over 102 weeks of follow-up, the placebo-subtracted proportional change in urinary albumin:creatinine ratio (UACR) was −47.1% (95% CI −64.8 to −20.6) with dapagliflozin 5 mg/day and −38.4% (95% CI −57.6 to −10.3) with dapagliflozin 10 mg/day [14]. Dapagliflozin 5 and 10 mg/day had no effect on glycaemic control; the placebo-subtracted change in haemoglobin A1c (HbA1c) was only 0.03% (95% CI −0.3–0.3). Similarly, compared with placebo, dapagliflozin 5 mg/day provoked a non-significant reduction of −1.4 mmHg in systolic blood pressure (BP) (95% CI −4.8–2.0) and dapagliflozin 10 mg/day provoked a modest reduction of −3.8 mmHg (95% CI −6.9 to −0.7) in systolic BP [14]. Adjustments for treatment-induced changes in HbA1c or systolic BP only modestly attenuated the reduction in UACR provoked by the SGLT-2 inhibitor.

Taken together, the above results suggest that the albuminuria-lowering effect of dapagliflozin that was evident in diabetic patients with Stages 3b–4 CKD cannot be explained by the glucose- or BP-lowering efficacy of this agent. It is not surprising that the level of eGFR modifies the effect of SGLT-2 inhibitors on glycaemic control [15]; the attenuated hypoglycaemic action of these agents is most likely attributable to the reduced filtration of glucose and this was rather clear in this study that included patients with baseline eGFR <45 mL/min/1.73 m². In addition, the modest BP-lowering effect of dapagliflozin in this combined analysis is in accordance with the results of a meta-analysis of seven randomized trials showing that the placebo-subtracted reduction in 24-h ambulatory BP provoked by SGLT-2 inhibitors is −3.62 mmHg for systolic and −1.70 mmHg for diastolic BP [16]. Incorporating data from randomized trials using the ‘gold standard’ method of ambulatory BP monitoring, this meta-analysis showed that the BP-lowering effect of SGLT-2 inhibitors is similar to that of low-dose hydrochlorothiazide [16]. Based on this meta-analytic evidence, the cardiorenal protection afforded by SGLT-2 inhibitors appears to be independent from haemodynamic benefits. Larger randomized trials evaluating ‘hard’ endpoints are now warranted to elucidate the safety and efficacy of SGLT-2 inhibition in patients with DM and eGFR <30 mL/min/1.73 m². Other trials such as EMPA-KIDNEY are testing the utility of SGLT-2 inhibitors among patients without DM.

ADEQUATE MANAGEMENT OF VOLUME OVERLOAD IN HAEMODIALYSIS

Setting the ‘ideal’ dry weight for each individual patient and maintaining long-term euvolaemia represent primary goals of dialysis care [17]. In the absence of validated methods/techniques to guide volume management, the assessment of dry weight continues to rely on clinical judgement [17]. High interdialytic weight gain (IDWG) is often considered as a simple clinical marker of volume excess and limiting IDWG represents a standardized strategy to facilitate dry weight achievement [17]. In a retrospective analysis of 38 614 incident haemodialysis patients, Hecking et al. [18] explored the association of IDWG with pre- and post-dialysis volume overload, using bioimpendance spectroscopy (BIS) as a more objective method to assess volume status. The investigators reasoned that IDWG is an imprecise proxy of volume status, since this marker may be influenced by the nutritional state and other confounding factors. In a longitudinal part of this study, the investigators also explored the prognostic significance of pre- and post-dialysis volume overload versus IDWG [18].

Bivariate correlation analysis confirmed the original research hypothesis showing a weak relationship of IDWG with BIS-derived volume overload that was modified by the timing of measurements (pre- or post-dialysis). There was a weak positive correlation of IDWG with pre-dialysis volume overload (Spearman’s coefficient 0.24) and a weak negative correlation of IDWG with post-dialysis volume overload (Spearman’s coefficient −0.18) [18]. In the longitudinal part of the study, 5640 deaths were recorded over a median observational period of 491 days. Using time-varying 1-month averages of BIS-derived volume overload as predictors, the adjusted risk of all-cause death was significantly higher in participants stratified into Quartiles 3 and 4 compared with the reference Quartile 2 of volume overload. In contrast, participants stratified into Quartile 1 of IDWG had a significantly higher risk of all-cause death than those in the reference Quartile 2 [18]. In a combined analysis, the highest mortality risk was observed in participants stratified into Quartile 4 of pre-dialysis volume overload and Quartile 1 of IDWG [hazard ratio (HR) 2.66 (95% CI 2.21–3.20)] [18].

These observations suggest that haemodialysis patients with the least IDWG and most volume overload carry the highest mortality risk; this mortality risk appears to be mediated through chronic volume overload (not low IDWG). This notion is further supported by the Dry-weight Reduction in Hypertensive Haemodialysis Patients (DRIP) trial [19]. In a post hoc analysis, 150 hypertensive haemodialysis patients participating in DRIP were stratified into quartiles according to the intradialytic relative plasma volume (RPV) slope at baseline. Participants in Quartile 4 (i.e. those with a flatter RPV slope who were more likely to be volume overloaded) had the lowest IDWG. In contrast, participants in Quartile 1 (i.e. those with a steeper RPV slope who were less likely to be volume overloaded) had the highest IDWG (P < 0.001 for the trend across quartiles of RPV slope) [19]. Most importantly, when dry weight was probed over the 8-week long course of the DRIP trial, participants stratified into Quartile 4 of RPV slope achieved the greatest dry weight reduction (−1.5 kg); this improvement in volume status was accompanied by the highest reduction in 44-h ambulatory systolic BP (−12.6 mmHg) [19]. These observations suggest that low IDWG is not always better in haemodialysis [20]. In contrast, low IDWG should raise the suspicion of volume overload, particularly if accompanied by other clinical indications, such as the presence of uncontrolled interdialytic hypertension. In this rather common scenario, probing of dry weight is an opportunity to correct volume overload and improve clinical outcomes.

Additional insights on the importance of volume overload are provided by another study from Dekker et al. [21], in which the interrelation of pre-dialysis BP and volume status with all-cause mortality was explored in 8883 patients from the international Monitoring Dialysis Outcome Initiative database. Several observational cohort studies have shown that pre-dialysis systolic BP exhibits a U- or J-shaped association with mortality [22–24]. This study showed that this ‘paradoxical’ inverse association of lower pre-dialysis BP with higher mortality risk is confounded by volume status [21]. In participants assessed to be volume overloaded before dialysis with the use of BIS, low pre-dialysis systolic BP <110 mmHg was associated with increased mortality risk. Similarly, in those with pre-dialysis volume depletion, pre-dialysis systolic BP <140 mmHg was associated with a higher risk of all-cause death. In contrast, among euvolaemic patients before dialysis, low pre-dialysis systolic BP <110 mmHg was associated with a survival benefit. The survival benefit was the highest among patients with slight volume depletion after dialysis (< −1.1 L) and post-dialysis systolic BP within the ‘normal’ range of 110–140 mmHg [HR 0.50 (95% CI 0.29–0.86)] [21]. Most importantly, when the risk association of ultrafiltration rate was analysed across subgroups of pre- or post-dialysis volume status, no significant association with all-cause mortality was evident [21]. Since adjusting for ultrafiltration rate did not modify the association of volume overload with mortality, recommendations based on limiting a maximum rate of ultrafiltration may not be grounded in physiology of dialysis [20]. What appears to be more important is the achievement of an adequately low post-dialysis weight. Proof that this strategy is beneficial needs to be provided by randomized trials in the future.

FUNDING

R.A. is supported by the National Institutes of Health (5 R01 HL126903-02) and a grant from Veterans Affairs Merit Review (5I01CX000829-04).

CONFLICT OF INTEREST STATEMENT

R.A. is a member of data safety monitoring committees for AstraZeneca and Ironwood Pharmaceuticals; a member of the steering committees of randomized trials for Akebia, Bayer, Janssen, GlaxoSmithKline, Relypsa, Sanofi and Genzyme; a member of adjudication committees for Bayer, Boehringer Ingelheim and Janssen; and a member of the scientific advisory board or consultant for Celgene, Daiichi Sankyo, Eli Lilly, Relypsa, Reata, Takeda Pharmaceuticals USA and ZS Pharma. P.I.G. has nothing to disclose.