Non-frozen transports of whole blood samples do not cause relevant bias for global coagulation tests in clinical trials evaluating the drug safety

Sample shipments with dry ice have a large economic impact on clinical research. Therefore, the bias caused for global coagulation tests by non-frozen transports of whole blood instead of frozen plasma was investigated experimentally and by a meta-analysis of 6-year central laboratory data. In the experiment, aliquots from 14 healthy volunteers were kept as whole blood at 20 ± 2 °C and as frozen plasma until an analysis of prothrombin time (PT), activated partial thromboplastin time (aPTT), thrombin time (TT), and antithrombin III (ATIII) at day 0, 1, 2, and 3 from collection. Within these 3 days only PT and aPTT demonstrated any changes: in blood samples kept at 20 ± 2 °C these amounted about 10% for both. In frozen plasma, aPTT did not change whereas PT increased by 14%. In a meta-analysis of central laboratory data, PT and aPTT results were grouped across various phase II-IV trials by the type of sample transfer, either as frozen plasma on dry ice or non-frozen as whole blood. For the latter the mean difference to a reference group of phase I trials with same-day analysis was in line with the amount of bias found in the experiment (aPTT, 34.6 ± 6.0 vs. 31.6 ± 3.5 s; PT, 87.7 ± 13.3 vs. 97.3 ± 7.9%). The consistent bias resulted in shifted, but still normal distribution curves with a total rate of clinically relevant outliers of about 1.9% for aPTT and 2.4% for PT. Biases thus appear irrelevant for a common safety evaluation within clinical trials. Non-frozen whole blood transports for the measurement of global coagulation tests appear justified for this purpose, if protocols do not require frozen shipments for other reasons. However, transit time must not exceed 2 days and pre-analytical conditions should be consistent within the same trial.