Reproducibility of blood tests of liver fibrosis in clinical practice.
Calès P, Veillon P, Konaté A, Mathieu E, Ternisien C, Chevailler A, Godon A, Gallois Y, Joubaud F, Hubert-Fouchard I, Oberti F, Réaud S, Hunault G, Mauriat F, Lunel-Fabiani F
2008 Clin. Biochem. Volume 41 Issue 1-2
To evaluate the inter-laboratory reproducibility of blood test for liver fibrosis: FibroMeter, Fibrotest, APRI and their composites variables.
DESIGN AND METHODS
Four studies, including 147 patients, were performed: study #1 included 2 metachronous blood samples and 2 laboratories; studies #2, #3 and #4 included synchronous samples with assays delayed at day 1 in 12 laboratories, at day 0 in 10 laboratories and at day 0 or 1 in 2 laboratories, respectively. Agreement was evaluated by the intraclass correlation coefficient (r(ic)).
In studies #1, #2 and #4, r(ic) for FibroMeter was 0.893, 0.942 and 0.991, respectively. In study #3, the r(ic) were: FibroMeter: 0.963, Fibrotest: 0.984, APRI: 0.949. Large simulated variations in composite variables had a weak impact on FibroMeter.
When blood marker limits are controlled, inter-laboratory agreement of blood tests is excellent in clinical practice conditions. Blood tests are robust against the variability of composite blood variables.