The accurate standardization of the preanalytical phase is of pivotal importance for achieving reliable results of coagulation tests. Because information on the suitable storage conditions for coagulation testing is controversial, we aimed at investigating the sample stability with regard to the temperature and time before centrifugation. The activated partial thromboplastin time (aPTT), prothrombin time (PT), fibrinogen and D-dimer were assayed in specimens collected from 26 consecutive patients on antivitamin K therapy on the ACL TOP analyzer. Three primary 3.6-ml siliconized evacuated tubes containing 0.109 mol/l buffered trisodium citrate were sequentially collected from each patient. These three tubes were mixed, pooled and divided into seven identical aliquots. The first aliquot was immediately centrifuged according to the standard protocol [1500 g for 15 min at room temperature (RT)] and analyzed. The other aliquots were left for 3, 6 and 24 h, respectively, at RT or 4 degrees C, and then centrifuged and analyzed. Test results were compared with those obtained on the reference specimen. Statistically significant prolongations were observed for aPTT in all the samples. Such differences exceeded the analytical quality specifications for desirable bias in the samples stored for 24 h. A significant reduction, yet comprised within the desirable bias, was observed for PT and fibrinogen in uncentrifuged specimens stored at RT for 3 and 6 h. No significant biases could be recorded in D-dimer. In conclusion, a 6-h storage of uncentrifuged specimens at either RT or 4 degrees C may still be suitable to achieve results of routine coagulation testing comprised within the analytical quality specifications for desirable bias.