Graphical Abstract
Top panel: meta-analysis (random effects model) for occurrence of end-stage kidney disease in four major outcomes trials of SGLT2 inhibitors in heart failure. Bottom panel: depiction of different mechanisms driving the progression of chronic kidney disease in type 2 diabetes, as compared with chronic heart failure. Drugs that lower intraglomerular pressures may be beneficial in type 2 diabetes (characterized by glomerular hyperfiltration), but detrimental in chronic heart failure (characterized by glomerular hypofiltration).
In large-scale randomized controlled trials, several classes of drugs have been shown to reduce major adverse renal outcomes in Type 2 diabetes. As compared with placebo, angiotensin receptor blockers, angiotensin converting enzyme (ACE) inhibitors, mineralocorticoid receptor antagonists, and sodium–glucose cotransporter 2 (SGLT2) inhibitors each reduce the risk of occurrence of end-stage kidney disease by ≈30%–35%.[1–3] Therefore, many of the foundational drugs that favourably influence cardiovascular death and heart failure hospitalization in chronic heart failure act to decrease the risk of serious adverse renal outcomes in diabetic patients, most of whom did not have heart failure.
Yet, surprisingly, large-scale trials have not demonstrated that these four classes of drugs reduce the risk of end-stage kidney disease in patients with heart failure, even though the same trials demonstrated significant benefits of these treatments on major heart failure outcomes. Long-term therapy with angiotensin receptor blockers and ACE inhibitors in patients with heart failure has not altered the rate of decline in glomerular function or reduced the risk of developing severe renal impairment, and initiation of treatment with these drugs increases the likelihood that patients will experience a decline in estimated glomerular filtration (eGFR) to <30 mL/min/1.73 m2 nearly five-fold,[4] even when therapy is started at low doses. Similarly, treatment with mineralocorticoid receptor antagonists has not influenced the progression of chronic kidney disease in patients with heart failure, and initiation of treatment with spironolactone has increased the likelihood of worsening kidney function by >60% in patients with a preserved ejection fraction and more than two-fold in those with a reduced ejection fraction.[5,6] Yet, worsening of kidney function with ACE inhibitors and mineralocorticoid receptor antagonists, although common, did not negate the benefits of these drugs to reduce the risk of major heart failure events.
The totality of evidence does not even support a meaningful renoprotective effect of SGLT2 inhibitors in patients with heart failure. Empagliflozin and dapagliflozin have been reported to favourably influence the slope of decline in eGFR; however, the calculation of eGFR slopes in these trials are artificially flattened by the exclusion of changes during the first month of treatment, during which patients commonly experience a modest increase in serum creatinine. Yet, the most persuasive renal outcome is not eGFR slope, but the development of end-stage kidney disease, defined as renal replacement therapy or an estimated glomerular filtration rate <10–15 mL/min/1.73 m2. When the four large-scale trials of SGLT2 inhibitors in heart failure are combined in a meta-analysis (random-effects model), there is little evidence for a favourable effect of SGLT2 inhibitors on the development of end-stage kidney disease [hazard ratio 0.88 (0.61–1.26)], Figure 1. This finding stands in contrast with the 35% reduction in end-stage kidney disease noted in large-scale randomized trials of SGLT2 inhibitors in patients with type 2 diabetes [hazard ratio 0.65 (0.53–0.81)].[3] Even when the evaluation of renal outcomes is broadened to include patients with a sustained ≥50% decrease in eGFR from baseline, three of the four major heart failure trials with SGLT2 inhibitors (EMPEROR-Preserved, DAPA-HF, and DELIVER) failed to show a benefit of these drugs, based on a total of ≈300 events.[7] In the EMPA-KIDNEY trial, the hazard ratio for the reduction in the risk of a major renal outcome with empagliflozin (using the broadened definition of a renal event) was 0.70 [95% confidence interval (CI): 0.61–0.80] in patients without heart failure and 1.00 (95% CI: 0.67–1.47) in patients with heart failure.[8] The lack of a renal benefit in heart failure cannot be explained by differences in the baseline glomerular filtration rate, since the patients in the SGLT2-inhibitor heart failure trials had a glomerular filtration rate in between those with Type 2 diabetes and those with chronic kidney disease, both of whom experienced meaningful nephroprotection with SGLT2 inhibition.
Figure 1.
Effect of SGLT2 inhibitors on the occurrence of end-stage kidney disease in large-scale trials of patients with heart failure, and contrasting mechanisms of progressive of kidney disease in Type 2 diabetes and chronic heart failure. ACE, angiotensin converting enzyme; SGLT2, sodium–glucose cotransporter 2.
What might explain differences in the results with angiotensin receptor blockers, ACE inhibitors, mineralocorticoid receptor antagonists, and SGLT2 inhibitors in patients with Type 2 diabetes vs. those with heart failure? One possibility is that cardiovascular death and heart failure hospitalizations are 15–20 times more common than renal events in patients with heart failure, and the occurrence of a major heart failure event might represent a competing risk for the occurrence of a major renal event. At least theoretically, the prevention of heart failure events might increase the likelihood that physicians would identify the occurrence of a marked worsening of kidney function as a renal event (rather than a heart failure event), thus making patients in the active treatment arm more vulnerable to the apparent occurrence of attributable renal events and potentially biasing the estimates for any effect of treatment on renal events toward the null. However, it should be noted that, in large-scale trials of patients with type 2 diabetes, a heightened ratio of heart failure events to end-stage kidney disease is also observed and worsening heart failure events are prevented by active treatment, and yet, angiotensin receptor blockers, ACE inhibitors, mineralo-corticoid receptor antagonists, and SGLT2 inhibitors exert a striking renoprotective effect in diabetic patients.
A second possibility may lie in the duration of the studies. The trials of SGLT2 inhibitors in heart failure were often shorter, being driven to completion by the inclusion of nonfatal heart failure hospitalizations in the primary endpoint analysis. The EMPEROR-Reduced and DAPA-HF trials had a median duration of 16–18 months, and the Kaplan–Meier curves for the occurrence of renal events in most large-scale trials in Type 2 diabetes did not diverge during the first year.[2,7] If a favourable effect on eGFR slopes were to be translated into a reduced risk of renal events, a significant benefit might require longer periods of active treatment. However, it should be noted that the EMPEROR-Preserved and DELIVER trials showed no difference in renal outcomes in trials with a median duration > 2 years, a duration comparable to that of the trials of SGLT2 inhibitors in type 2 diabetes, and there was no indication of separation of the Kaplan–Meier curves at later timepoints. Similarly, in EMPA-KIDNEY, the effect of empagliflozin on renal events differed in patients with and without heart failure, even though the two groups had the same duration of follow-up (median 2 years).[8] Of note, the median duration of the trials of ACE inhibitors, angiotensin receptor blockers, and mineralo-corticoid receptor antagonists exceeded 2–3 years, a duration which should have been sufficient to discern a benefit on kidney outcomes if it was present.
A third possibility is that the mechanisms that drive the progression of kidney disease in patients with heart failure may differ from those in diabetes (Figure 1). Glomerular hyperfiltration (perhaps due to hyperglycaemia-related proximal tubular sodium reabsorption) appears to drive the loss of nephrons in diabetic kidney disease,[9] and the renoprotective effects of conventional inhibitors of the renin–angiotensin system, mineralocorticoid receptor antagonists, and SGLT2 inhibitors are commonly attributed to their action to alleviate heightened intraglomerular filtration pressures. In contrast, glomerular hyperfiltration has not been demonstrated in patients with heart failure, and instead, when decreases in renal arterial perfusion are coupled with increases in renal venous pressure, glomerular filtration pressures are likely to be severely compromised, and efferent arteriolar vasoconstriction may be essential to maintaining glomerular function.[10–12] Therefore, under conditions where renal venous pressures are elevated and glomerular hypofiltration may be present, drugs that cause efferent arteriolar vasodilation (e.g. conventional inhibitors of the renin–angiotensin system, mineralocorticoid receptor antagonists, and SGLT2 inhibitors) may further lower intraglomerular pressures, potentially to critically important levels. It is therefore noteworthy that the marked renal decongestion that is seen following implantation of a left ventricular assist device can produce remarkable short-term improvement in renal function in patients with advanced heart failure.[13]
This conceptual model suggests thatin a congested kidney in which efferent arteriolar vasoconstrictive support mechanisms are paralyzed by conventional inhibitors of the renin–angiotensin systemafferent arteriolar tone may be a critical determinant of glomerular function. Indeed, as a result of its ability to potentiate the action of endogenous natriuretic peptides to cause afferent arteriolar vasodilation, neprilysin inhibition enhances glomerular function in heart failure.[14] This may explain the effect of sacubitril/valsartan (when compared with conventional inhibitors of the renin–angiotensin system) to reduce the risk of major adverse kidney outcomes in patients with heart failure, even though the drug acts to increase albuminuria.[14,15]
These observations, taken together, lead to three important conclusions. First, end-stage kidney disease is a distinctly uncommon event in trials of patients with chronic heart failure trials. When the four SGLT2-inhibitor trials in heart failure are combined, there were > 1800 cardiovascular deaths, but only 120 occurrences of end-stage kidney disease. The contribution of renal events to major health outcomes would be even smaller if nonfatal hospitalizations for heart failure were included in the analysis of major heart failure outcomes. Second, many drugs that reduce cardiovascular death in patients with heart failure are accompanied by worsening glomerular function, which does not negate the benefits of treatment.[5–8] Third, the simultaneous occurrence of renal hypoperfusion and congestion—rather than intraglomerular hypertension—may be the primary factor that drives the development of end-stage kidney disease in patients with heart failure.
Data availability
The paper performed a meta-analysis using an open source program. All data used in this analysis are already fully presented in the figure.
Funding
All authors declare no funding for this contribution.
Eur Heart J. 2023;44(17):1522-1525. © 2023 Oxford University Press
Copyright 2007 European Society of Cardiology. Published by Oxford University Press. All rights reserved.
