Probing the Limits of Permissive Hypercreatininemia during Decongestion in Patients with Heart Failure

When managing patients with heart failure, few clinical decisions provoke more consternation than the rise in serum creatinine after diuretic intensification or employment of the multiple pillars of guideline-directed medical therapies (GDMTs). For decades, this so-called worsening kidney function has been viewed with suspicion, often prompting premature halting of therapy despite well-established benefits of volume removal and neurohormonal blockade.¹ In recent years, however, data from several studies have consistently reinforced a counter-narrative that mild-to-moderate reductions in GFR, particularly when occurring in the context of effective decongestion in the acute phase² or the application of GDMT in the outpatient setting,³ may represent a tolerable—if not expected—physiologic response rather than a signal of kidney injury or harm.⁴ Importantly, nearly all studies demonstrated better outcomes regardless of the degree of GFR decrement or the achieved absolute eGFR after decongestion or therapy initiation, with no observable threshold at which the kidney signal trumped the salutary benefits of the heart failure therapies in either the acute² or chronic setting.⁵

These findings consistently reinforced a growing paradigm that myself and others have embraced as a benign form of AKI that should be dismissed as permissive hypercreatininemia, particularly in patients for whom volume overload is a key driver of symptoms and risk.^5–7 Even post hoc analyses of randomized controlled trials that ascertained several urinary biomarkers of kidney injury revealed that even in those patients with heart failure with probable tubular injury, in addition to decreases in kidney function, there was still no signal for harm in terms of hard clinical outcomes.^4,8 Detailed examination revealed that those who had higher concentrations of biomarkers of kidney injury had the most fluid removed and most decongested in the acute phase of heart failure.^4,8

Thus, one could argue that we need not heed much attention to the kidney function, so long as the clinical parameters were tracking in the right direction, unless reaching an extreme level of kidney dysfunction.⁹ But how much is too much kidney dysfunction? How much change in eGFR is too much change, how much absolute kidney function can be tolerated until the risk for poor outcomes is in fact increased and clinical management should be altered? I generally would advise to others that those lines in the sand can only be determined on a case-by-case basis by employing the art of medicine.

In this issue of CJASN, Oka et al. push our knowledge boundaries on the interplay between the kidneys and heart failure in the postacute setting with a sophisticated post hoc analysis of the Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study With Tolvaptan trial population, evaluating dynamic changes in both eGFR and clinical congestion score among >3400 patients discharged after hospitalization for acute heart failure.¹⁰ Using marginal structural models to account for time-varying confounding and immortal time bias, the authors classified patients into eight exposure groups based on whether they exhibited improved or worsened congestion in combination with degrees of eGFR decline (none, 1%–20%, 21%–40%, ≥41%). Patients with postdischarge improvements in congestion had significantly lower mortality—even when accompanied by up to a 40% decline in eGFR—compared with those with persistent congestion and preserved kidney function. However, when the eGFR decline exceeded 40%, decongestion no longer conferred a protective advantage, and the risk for death was approximately two-fold higher compared with worsened congestion and no eGFR decline.¹⁰

Oka et al. are to be commended for their methodologic rigor, leveraging repeated measures of both kidney function and clinical congestion status and inverse probability weighting and marginal structural models to maximize the opportunity to glean causal inference from observational data. Visits were every 8 weeks, which minimizes the chances for reverse causality, which can be potentially imagined in the scenario where the clinicians see the dropping eGFR and withhold diuretics in the interim, and at the next visit, the patient has become more congested, and due to the effects of congestion on kidney function, have even worse kidney function. These potential biases were also minimized by the elegant analytic design.¹⁰

Importantly, the Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study With Tolvaptan cohort offers a unique vantage point from the prior studies in this vein. Unlike in-hospital heart failure studies where the changes in kidney function and achievement of decongestion occurred during the boundaries of the hospital admission, this analysis extends the ascertainment window into the chronic phase of heart failure. Thus, the most relevant comparable to this study are post hoc analyses of changes in eGFR chronic angiotensin converting enzyme inhibitor therapy,³ spironolactone,¹¹ sodium-glucose cotransporter-2 inhibitors,⁵ or sacubitril/valsartan.¹² This study is consistent with those data and suggests that eGFR decline may be clinically acceptable when linked to meaningful improvements in volume status or application of GDMT. This study tells us for the first time, however, where the demarcation line in terms of kidney function that we may be pushing decongestive therapies too far. In this cohort, at the least, when the decline in eGFR exceeded 40%, even improved congestion may not overcome the associated hazard with such a large decrement in eGFR. We must also acknowledge the unmeasured confounding that may be present. The study's reliance on clinical congestion score and B-type natriuretic peptide to define congestion, while pragmatic, cannot fully capture intravascular versus interstitial volume or the complex interplay of right heart pressures and kidney perfusion. Nor can eGFR alone distinguish between functional hemodynamic perturbations and true structural renal injury. In fact, there is also likely a subgroup of patients who had worsening of kidney function that was independent of the cardiac function and hemodynamics. Nonetheless, these data move us closer to an actionable framework—one in which trajectory matters more than static eGFR values, and context trumps nephrocentrism.

As clinicians, we often fall victim to a kind of creatinine-phobia—an overreaction to changes in serum creatinine that leads to deimplementation of life-saving therapy. Instead, we should have more of a zen-like approach in which we expect and tolerate some degree of renal perturbation when using effective therapies in heart failure. Our task is not to avoid all changes in creatinine but to discern when those changes are meaningful and require a change in actions. To that end, a practical recommendation is to expect and tolerate eGFR dips up to approximately 30%–40% in the context of improved congestion or initiation of GDMT (Figure 1). In this regard, nearly 50% of patients who were decongested had an eGFR decline of up to 40% within the 44 weeks. By contrast, the large 40% decline in eGFR was rare—only 2% of decongested patients experienced this magnitude of eGFR decline. Although these data need validation in contemporary cohorts and in patients on modern GDMT, the large decline in eGFR within a year aligns nicely with the US Food and Drug Administration/National Kidney Foundation endorsed clinical kidney end point of 40% decline in eGFR for long-term kidney disease trials as a strong surrogate for future ESKD.¹³

Figure 1.

The interplay between heart failure therapies and kidney function with benefits versus risk across the spectrum of eGFR decline. GDMT, guideline-directed medical therapy.

As the great Richard Feynman once said, “We are trying to prove ourselves wrong as quickly as possible, because only in that way can we find progress.” For too long, we have clung to the idea that any rise in creatinine signals harm. The recent studies have shown us how important clinical context is for determining the significance of these changes. Studies like this help us define the boundaries of our newer concepts—and move toward a more enlightened, evidence-informed approach to cardiorenal care.

Supplementary Material

cjasn-20-1174-s001.pdf ^{(1.4MB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/CJN/C356.

Author Contributions

Conceptualization: Steven G. Coca.

Writing – original draft: Steven G. Coca.

Writing – review & editing: Steven G. Coca.

Funding

None.