In this issue of the Journal, Smith and colleagues report their findings that demonstrate increases in serum creatinine levels during heart failure hospitalization are associated with worsened outcomes after hospital discharge. The authors conducted a cohort study of 412 patients hospitalized at their institution over a 30-month period who survived to discharge. The change in serum creatinine level (ΔCr) was defined simply as the difference between the baseline and discharge levels, with “worsened renal function” defined as a ΔCr ≥ 0.1mg/dL. The outcomes examined in this study were survival, readmission, and change in functional status during the 6 months after hospitalization. A ΔCr ≥ 0.3mg/dL was associated with a 60% increased adjusted mortality risk, and surprisingly the ΔCr was a stronger mortality predictor than the baseline creatinine level. Furthermore, the higher mortality risk associated with ΔCr ≥ 0.3mg/dL was similar at all levels of baseline creatinine level.
These findings build upon prior studies that have determined the importance of renal function in the prognosis and management of heart failure. Most previous analyses evaluating the effect of renal function on heart failure outcomes have used a baseline measure from the outpatient setting.1 A prior study from this group of authors similarly found elevations of creatinine ≥ 0.3 mg/dL during heart failure hospitalization to be associated with greater in-hospital length of stay and mortality risk. The current study extends these findings by demonstrating an association of increased ΔCr with worse outcomes for survivors of hospitalization up to 6 months after discharge, and showing a dose-dependent relationship of ΔCr and increased mortality risk. However, as with most interesting studies, this analysis provokes as many questions as it answers for clinicians managing heart failure patients. The primary questions that arise from this study are: (1) Can the ΔCr be interpreted as a “diagnostic test” for high mortality risk? (2) Is worsened renal function a causal mechanism leading to increased mortality risk or a marker of poor cardiac function and subsequent under-treatment? and (3) What are the clinical implications of these findings for inpatient and postdischarge heart failure management?
In observational studies, it is generally agreed that a relative risk of 2.0 or greater demonstrates a strong association between the predictor and the outcome—in this case ΔCr and mortality after hospitalization. However, if we use the presence or absence of a risk factor to alter our course of therapy, then we have transformed the risk factor into a diagnostic test. Because diagnostic tests are used to determine the use or nonuse of potentially important treatments, we hold them to a higher standard than risk factors. If a clinician changes the plan of therapy in heart failure because of an increased ΔCr, then worsening renal function effectively becomes a diagnostic test. The optimal method for evaluating a diagnostic test is by determining the positive and negative likelihood ratios (+LR and −LR) at the cut point of interest; a +LR > 5.0 and a −LR < 0.5 have been described as minimal thresholds for a useful diagnostic test.2 In Table 1 of the Smith study, ΔCr ≥ 0.5 mg/dL has an adjusted relative risk of 2.9, which is a substantial increased risk for mortality. Nonetheless, the +LR (sensitivity/1-specificity) is only 1.6 and the −LR (1-sensitivity/specificity) is 0.8. These results, which are clearly presented by the authors, instruct us statistically that we should not use the ΔCr as a diagnostic test per se, despite its strong association with mortality.
The second important question that arises from this study is why increased serum creatinine concentrations portend such an adverse prognosis, even after statistical adjustment for age, diabetes, ejection fraction, baseline creatinine, and duration of heart failure? One possibility is that a reduction in renal function over a short period spurs the progression toward poor outcomes because renal dysfunction directly contributes to heart failure risk. Numerous physiologic perturbations accompany renal dysfunction, such as volume retention, decreased hemoglobin, increased inflammatory factors, homocysteine levels, and asymmetric dimethyl arginine levels (a marker of endothelial dysfunction). Adverse drug events are also more common in persons with renal dysfunction, in particular reactions to heart failure agents including renin angiotensin aldosterone (RAA) antagonists, spironolactone, and digoxin. An incremental drop in renal function could theoretically reduce survival via the above mechanisms.
We believe the more likely explanation is that increased creatinine levels during hospitalization are a marker of poor cardiac output, leading to diminished renal blood flow and reduced ability to tolerate inpatient heart failure treatment. Some have described the azotemia that occurs with reductions in preload and systemic vascular resistance as “cardiorenal” syndrome. The authors accounted somewhat for severity of heart failure by adjusting for the left ventricular ejection fraction, but increased creatinine during aggressive diuresis is probably another measure of poor cardiac reserve. Thus, the change in creatinine is more likely to be indicative of the severity of cardiac dysfunction rather than of acute renal damage. Indeed, one of the most effective ways to mitigate rises in creatinine is the transient administration of positive inotropic agents, although this approach (or possibly the need for inotropic agents) is also associated with a poor prognosis.
Another potential reason for adverse heart failure outcomes among patients with renal dysfunction may be the reduction or withholding of certain beneficial therapies. In fact, serum creatinine levels often decline when RAA inhibitors are withdrawn or diuretics withheld, although these actions will likely worsen prognosis and increase the chance of early readmission. Elevated or increasing serum creatinine levels have been associated with lower use of both angiotensin-converting enzyme (ACE) inhibitors and β-blockers, the agents that have the greatest impact on post-discharge outcomes.3
This leads us to the third question of whether increased creatinine levels should alter patient management in hospitalized heart failure patients. In most cases, the answer should be an emphatic “no.” As outlined previously, the increased creatinine level is usually a marker of poor cardiac output, so the ΔCr is not the disease to be treated. Reducing the dose or discontinuing the ACE inhibitor or diuretic might make the marker look better, but will be unlikely to help the patient. Inadequate diuresis at discharge is the most significant indicator of early readmission. Although Smith et al evaluated inpatients, the experience with rising creatinine during outpatient therapy is instructive. Outpatients with heart failure initiated on ACE inhibitors will often have transient increases in the creatinine level. The Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS) trial found ACE-inhibitors to be at least as effective in those subjects with substantially increased creatinine levels (>30%) than in subjects without increased creatinine.4 Small increases in the serum creatinine level with the initiation of ACE inhibitors should not be an indication for discontinuing therapy, and indeed may be a marker for patients who would derive the greatest benefit.
What this study does tell us is that we need to avoid nonessential agents that have the propensity to worsen renal function, such as nonsteroidal anti-inflammatory drugs and, in some patients with low blood pressure (and possibly occult renal artery stenosis), excessive use of direct vasodilators. In addition, there is a great need for heart failure therapies that are renal-protective and that can be used in patients with severe renal dysfunction. Currently, there is little evidence to guide the treatment of this high-risk subgroup of heart failure patients, because they are systematically excluded from most heart failure clinical trials. Whether preventing further increases in the serum creatinine level during hospitalization can lead to improved outcomes in heart failure is unknown. If so, the benefit will come from new pharmaceutic agents that preserve renal blood flow and prevent or treat the so-called cardiorenal syndrome. A beneficial effect of preventing worsening renal function will not derive from withholding life-extending therapies out of fear of the rising creatinine level, and given current usage patterns, it is likely that this practice contributes to the poor outcomes of these patients.
FULL TEXT