Format

Send to

Choose Destination
Clin Nephrol. 2004 Nov;62(5):344-50.

Evolution and predictive power of serum cystatin C in acute renal failure.

Author information

1
Department of Surgery, Division of Anesthesiology and Intensive Care Medicine, Intensive Care Unit, Helsinki University Hospital, Helsinki, Finland. annika.ahlstrom@hus.fi

Abstract

AIMS:

The serum concentration of cystatin C has recently been proposed as a better indicator of glomerular filtration rate (GFR) than plasma creatinine. Little is known about cystatin C in critical illness. We assessed serum cystatin C as a marker of renal function in acute renal failure (ARF) and its power in predicting survival of ARF patients.

MATERIAL:

202 consecutive adult patients admitted into the intensive care unit (ICU) during a period of 9 months.

METHOD:

Serum cystatin C, plasma creatinine and plasma urea were measured on admission, daily during the first 3 days, and 5-7 times a week during the rest of the ICU stay. The patients with and without ARF were compared by the Mann-Whitney U-test. The correlation between different variables was calculated by Spearman's correlation. Forward stepwise multiple regression analysis was performed to test independent predictors of mortality. The positive predictive value of serum cystatin C and plasma creatinine for ARF and mortality was calculated by ROC analysis.

RESULTS:

ARF occurred in 54 patients (27%). Serum cystatin C showed excellent positive predictive value for ARF in critical illness by ROC analysis. In acute renal dysfunction, abnormal values of serum cystatin C and plasma creatinine appeared equally quickly (median 3 days). The diagnosis of ARF, the day 1 Apache II score and admission plasma creatinine appeared as independent predictors of hospital mortality. ROC analysis showed only weak predictive power for serum cystatin C and plasma creatinine regarding hospital mortality.

CONCLUSIONS:

Serum cystatin C was as good as plasma creatinine in detecting ARF in intensive care patients. Neither marker was clinically useful in predicting mortality.

PMID:
15571178
DOI:
10.5414/cnp62344
[Indexed for MEDLINE]

Supplemental Content

Loading ...
Support Center