Skip to main content


Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

Are serum cystatin-C-based estimates better than those derived from serum creatinine in critically ill patients?

  • 650 Accesses

  • 1 Citations


The aim of our study was to evaluate the utility of two cystatin-C-based equations, as a surrogate of the renal function (glomerular filtration rate (GFR)) in a group of critically ill patients.


This was a monocentric, prospective and observational study including 146 samples respecting 22 ICU patients. Daily evaluation of seric creatinine, seric cystatin C (CC) and 24-hour creatinine clearance (24CrCL) was performed during the ICU stay. Comparisons were done between two CC-based equations (Hoek (H) and Larsson (L) formulas) and: 24CrCL; Cockroft-Gault (CG); modified Cockroft-Gault (mCG); and six-variable Modification of Renal Disease (MDRD6) formulas. Patients with chronic renal failure were excluded. Correlation, precision, bias and discrimination power were assessed using Spearman coefficient, Bland-Altman plots and receiver operating characteristic curves.


The average age of the patients was 63.4 years and male gender was predominant (68%). The APACHE II score was 16.8 ± 5.7. The medians of H and L estimates were 50.5 (28 to 77.6) and 47.7 (24.5 to 79.2) respectively, as compared to 69.8 (29.8 to 115.7), 60.7 (42.6 to 101.4), 58.9 (42.6 to 65.1) and 59.2 (40.6 to 106.8) ml/minute/1.73 m2, respectively to 24CrCL (reference method), CG, mCG and MDRD6. Correlation (r) between H, F, CG, mCG, MDRD6 and 24CrCL was 0.83/0.83/0.73/0.70/0.74, respectively. H and L formulas showed the smallest bias and limits of agreement, when compared with formulas based on serum creatinine, respectively -17.5/±52 ml/minute/1.73 m2 and -21/±52.8 ml/minute/1.73 m2. The sensibility for the identification of acute renal dysfunction (24CrCL under 60 ml/minute/1.73 m2) was high for H and L formulas (area under the curve of 0.94 for both). In the subgroup of 29 samples with 24CrCL above 130 ml/minute/1.73 m2 (patients with hyperfiltration) these two formulas had low sensibility (between 8 and 22%) for identification of this condition.


In this population of critically ill patients, cystatin-C-derived Hoek and Larsson equations underestimated 24CrCL; however, they have a better performance than the classic estimates (CG and MDRD6). Nevertheless, they are inaccurate when applied to ICU patients with hyperfiltration (GFR >130 ml/m/1.73 m2).

Author information

Correspondence to JP Baptista.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Baptista, J., Teixeira, S. & Pimentel, J. Are serum cystatin-C-based estimates better than those derived from serum creatinine in critically ill patients?. Crit Care 16, P356 (2012).

Download citation


  • Serum Creatinine
  • Glomerular Filtration Rate
  • Creatinine Clearance
  • Chronic Renal Failure
  • Characteristic Curf