Non-invasive estimation of glomerular filtration rate (GFR). The Lund model: Simultaneous use of cystatin C- and creatinine-based GFR-prediction equations, clinical data and an internal quality check.

Abstract Knowledge of glomerular filtration rate (GFR) is required to detect and follow impairment of renal function, to allow correct dosage of drugs cleared by the kidneys, and for the use of nephrotoxic contrast media. Correct determination of GFR requires invasive techniques, which are expensive, slow and not risk-free. Therefore, GFR-prediction equations based solely upon cystatin C or creatinine and anthropometric data or upon cystatin C, creatinine and anthropometric data have been developed. The combined prediction equations display the best diagnostic performance, but in several easily identifiable clinical situations (e.g. abnormal muscle mass, treatment with large doses of glucocorticoids) prediction equations based upon either cystatin C or creatinine are better than the combined equations. In Lund, where cystatin C has been used as a GFR-marker in the clinical routine since 1994, a strategy based upon this knowledge has therefore been developed. This comprises simultaneous use of a cystatin C-based and a creatinine-based GFR-prediction equation. If the GFRs predicted agree, the mean value is used as a reliable GFR-estimate. If the GFRs predicted do not agree, clinical data is evaluated to identify reasons for not using one of the two prediction equations and the GFR predicted by the other one is used. If no reasons for the difference in predicted GFRs are found, an invasive gold standard determination of GFR is performed. If the GFRs predicted agree for a patient, the creatinine value is reliably connected to a specific GFR and can be used to follow changes in GFR of that patient