Clinical Investigators
The prognostic importance of different definitions of worsening renal function in congestive heart failure*,**,

https://doi.org/10.1054/jcaf.2002.125289Get rights and content

Abstract

Background: Worsening renal function in patients hospitalized for heart failure portends a poor prognosis. However, criteria used to define worsening renal function are arbitrary, and the implications of different definitions remain unclear. We therefore compared the prognostic importance of various definitions of worsening renal function in 1,002 patients hospitalized for congestive heart failure (CHF). Methods and Results: The patient population was 49% female, aged 67 ± 15 years. Twenty-three percent had a prior history of renal failure, 73% had known depressed ejection fraction, and 63% had known CHF. On admission to the hospital, 47% were receiving ACE inhibitors, 22% β-blockers, 70% diuretics and 6% NAID's. 72% developed increased serum creatinine during the hospitalization, with 20% developing an increase of ≥ 0.5 mg/dL. Worsening renal function predicted both in-hospital mortality and length of stay > 10 days. Even an increased creatinine of 0.1 mg/dL was associated with worse outcome. Sensitivity for death decreased from 92% to 65% as the threshold for increased creatinine was raised from 0.1 to 0.5 mg/dL, with specificity increasing from 28% to 81%. At a threshold of a 0.3 mg/dL increase, sensitivity was 81% and specificity was 62% for death and 64% and 65% for length of stay >10 days. Adding a requirement of final creatinine of ≥ 1.5 mg/dL improved specificity. Conclusions: This analysis demonstrates that any detectable decrease in renal function is associated with increased mortality and prolonged hospital stay. This suggests that therapeutic interventions which improve renal function might be beneficial.

Section snippets

Data source

We obtained inpatient medical records for a geographically diverse sample of heart failure patients hospitalized between July 1, 1997, and June 30, 1998, at 11 academic medical centers. Consecutive heart failure hospitalizations were identified using the following International Classification of Diseases principal discharge diagnoses codes: 428.0, 428.1, 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, and 404.93.

Four nurses experienced in critical care or emergency care

Frequency of worsening renal function

The baseline characteristics of the patients are described in Table 1.

. Characteristics of patients

CharacteristicRangeN (%)
Demographics
 Age (years)67 ± 15
  Female490 (49%)
  White538 (54%)
Ejection fraction (n)
 ≥ 55%164 (16%)
 ≥ 40% and < 55%392 (39%)
 ≥ 20% and < 40%212 (21%)
 < 20%131 (13%)
 Missing105 (10%)
Medical history
 Prior heart failure636 (63%)
 Prior renal failure230 (23%)
 Hypertension703 (70%)
 Noninsulin dependent diabetes206 (20%)
 Insulin dependent diabetes204 (20%)
 Stroke155 (15%)
 Peripheral vascular disease

Discussion

This study demonstrates the sensitivity and specificity of various definitions of worsening renal function for prolonged hospitalization and in-hospital mortality. Of course, the value that is optimal depends upon which question is being asked. However, all the definitions that we evaluated demonstrate that worsening renal function portends a poor prognosis.

The utility of ROC curves for evaluating diagnostic curves of an illness is now well accepted.10, 11 Showing the sensitivity and

Conclusions

In conclusion, any worsening of renal function predicts increased mortality and prolonged hospital stays in patients hospitalized for congestive heart failure. However, ROC curves can help identify the definitions that may be most useful. Using appropriate definitions, it is possible that therapeutic interventions can be targeted to those most at risk.

References (11)

There are more references available in the full text version of this article.

Cited by (506)

View all citing articles on Scopus
*

This study was funded by Biogen, Inc., Cambridge, Massachusetts.

**

All editorial decisions for this article, including the selection of reviewers, were made by a guest editor. This policy applies to all articles authored by University of California–San Francisco faculty and staff.

Reprint requests: Stephen S. Gottlieb, MD, Division of Cardiology, University of Maryland School of Medicine, 22 South Greene Street, Baltimore, MD 21201.

View full text