A collaborative study by the Working Group on Hemostasis and Thrombosis of the Italian Society of Clinical Biochemistry and Clinical Molecular Biology (SIBioC) on the interference of haemolysis on five routine blood coagulation tests by evaluation of 269 paired haemolysed/non-haemolysed samples

Abstract

IntroductionHaemolysis is the leading cause of sample rejection in laboratory haemostasis. Most studies focused on artificially haemolysed samples. The aim of this study was a prospective assessment of spontaneous haemolysis on haemostasis tests, by comparing results of haemolysed (H) versus new, non-haemolysed (NH) specimens, collected within 4hrs. As new coagulometers can identify interfering substances, visual assessment of haemolysis was also compared with instrumental haemolysis index and stratified in subclasses.Materials and methodsTwo hundred and sixty nine paired samples were collected and analysed using ACL TOP750-CTS (Instrumentation Laboratory, Bedford, USA), for prothrombin time (PT), activated partial thromboplastin time (aPTT), D-Dimer (DD), fibrinogen (Fib) and antithrombin (AT). Bias between H and NH was calculated and compared with the respective critical difference (CD).ResultsMean bias was - 0.1 s for PT (P = 0.057), - 1.1 s for aPTT (P < 0.001), 1025 ng/mL for DD (P < 0.001), - 0.04 g/L for Fib (P = 0.258) and 1.4% for AT (P = 0.013). Bias exceeding the CD varied according to the method, with larger differences for aPTT (36.1%) and DD (17.1%) and < 8% for PT, Fib and AT. No correlation emerged between free haemoglobin values and difference in haemostasis tests in H and NH samples for any tests. Moderate/severe haemolysis involved > 95% of samples. The agreement between visual assessment and instrumental evaluation of haemolysis was 0.62.ConclusionSpurious haemolysis deeply influences aPTT and DD, and to a lesser extent AT and Fib. Prothrombin time seems only slightly influenced, suggesting that PT can be accepted also in haemolysed samples. Although a good inter-observer correlation of haemolysis evaluation was found, the instrumental assessment of haemolysis seems recommendable.

Hemolysis interference in 10 coagulation assays on an instrument with viscosity‐based, chromogenic, and turbidimetric clot detection

“…This minor effect of hemolysis has also been observed on an ACL TOP instrument. 15 The FVIII method, performed with a BioPhen reagent, did not show any sensitivity to hemolysis in the studied range (≤15 g/L). That is in agreement with previous findings by Laga et al 20 Among the turbidimetric immunoassays, vWF seems to be unaffected by hemolysis.…”

“…The potential interference of hemolysis on PT and APTT assays has been the subject of several studies. 7,[13][14][15][16][17]27,28 However, different combination of reagents and instrumentations makes the comparison between these studies cumbersome. 1 Stago STA R instruments have a viscoelastic mechanical clot detection.…”

“…In some of the hemolysis studies in coagulation assays, a second sample has been requested when the initial sample from a patient was found to be hemolyzed. 14,15 Even though this is a smart study design and give more representative samples, that approach is time-consuming and it is difficult to cover the whole concentration range of interest that are needed to determine a cutoff concentration where the samples should be suppressed. That is probably is the reason why in vitro induced hemolysis is the most commonly used method to study the influence of hemolysis.…”

“…1,28 For APTT, several studies with photo-optical clot detection have concluded that hemolysis does affect the measurements. 15,16,27 Among the studies of APTT with mechanical clot detection, a significant shortening, 17 small insignificant shortening, 14,18 and a prolongation of clotting times 7 as well as no significant difference 28 have been reported. In the present study with mechanical clot detection, a small but clinically irrelevant shortening of APTT was observed.…”

“…Novelli et al 15 that requested a second replacement sample for the hemolyzed patient sample. However, the results from other studies on the influence of hemolysis in D-dimer assays diverge.…”