Preanalytics For Coagulation Testing: Pitfalls And Possible Impact

By François Depasse, Diagnostica Stago, France , 3 May 2017

Preanalytics is crucial for clinical laboratory test results. Preanalytics covers all steps between laboratory test order and sample analysis. Most errors occur during this key step.

In this paper, various variables that may impact sample characteristics or integrity, including patient’s status, sample collection, management and storage will be discussed.

Patient’s status:

One prominent question patients and clinical laboratory may have before drawing blood for a haemostasis workup relates to patient’s fasting.

Scarce data are available on the duration of fasting and there is no well described or standardized definition of fasting on national or international grounds. Furthermore, the nature of clot detection system, i.e. optical versus viscosity-based detection system (also called mechanical detection) may also influence the sensitivity of the assay to patient’s fasting, especially for what regards possible post-prandial lipemia. In a recent study, Lima-Oliveira et al. studied the influence of a standardized light meal (563 kCal) on various haemostasis assays, including APTT, INR, fibrinogen, antithrombin, protein C and protein S. They demonstrated that a light meal had a minimal impact on these assay results.

The preferred time for blood sampling is 7 – 9 am when possible. Patients should refrain from smoking during at least 30 minutes before blood sampling: smoking increases platelet aggregability and induces a procoagulant state due to increased fibrinogen and PAI-1 levels and decreased tPA and plasminogen levels. Consumption of caffeine within two hours before blood draw is discouraged because of its influence on fibrinolytic activity. Physical activity should also be avoided within two hours before blood draw is also to be avoided as it increases leukocyte and platelet counts and results in coagulation activation. Moreover, strenuous physical exercise favors platelet microparticle release and induces a transient procoagulant state. On the other hand, stress should be avoided as it results in an increase of acute phase proteins, especially von Willebrand factor, factor VIII and fibrinogen. Blood should preferably be drawn from patients who have rested for a short period of time (5 minutes).

Devices for specimen collection:

Glass or plastic evacuated tube with a non-activating surface containing the appropriate additive (preferably trisodium citrate 105 – 109 mM, alternatively 129 mM provided labs have standardized to one citrate concentration and established citrate concentration-specific reference ranges), directly connected to the needle should be used. Winged devices may be preferable in certain situation (babies, children, patients with small veins or requiring frequent venipuncture). Syringe is an alternative to evacuated tube systems: in that case, small volume syringes (< 20mL) with blood added slowly to the appropriate volume of anticoagulant within one minute of blood draw and specimen immediately and properly mixed is recommended. The preferred needle gauge is 19 – 22; smaller gauge may induce hemolysis. In case of vascular access device (VAD), components of blood collection system must be checked to avoid air leaks that would result in an incorrect volume and hemolysis. Collection through lines previously flushed with heparin must be avoided (flush the line with 5mL of saline and discard the first 5mL of blood or six dead volumes of VAD). In case of blood obtained from a normal saline lock, two dead spaces volume of catheter + extension must be discarded.

Blood / anticoagulant ratio:

Citrate volume may have to adapt depending on patient’s hematocrit. Otherwise, in patients with an elevated hematocrit, plasma can be artificially diluted; this results in increased clotting times. CLSI recommends that citrate volume should be adapted for hematocrits above 0.55 and provides a formula and an abacus for determining the citrate volume versus hematocrit. On the opposite, there are insufficient date to support citrate volume adjustment for low hematocrits.

Specimen labelling:

As a general rule, specimen management should respect patient privacy. Patient and patient’s specimen should be positively identified at the time of collection labeled in the patient’s presence after the blood is collected The label should contain patient’s full name, unique identification number, date and time of collection, name of the person collecting the specimen, specimen type if a secondary or aliquot tube is used (anticoagulant type versus serum) and assay(s) performed (optional). An information request form may be transmitted to the lab along with the specimen.

Discard tube and order of blood draw:

PT (INR) and APTT results not adversely affected if tested on the first tube drawn, without discard tube. Current published data to support the asumption that a discard tube is necessary or unnecessary when drawing samples using a standard evacuated tube system for other tests are not available. When using a winged blood collection set and coagulation tube is the first tube drawn, draw a discard tube first (fill the blood collection tubing dead space, ensure maintenance of the proper blood / anticoagulant ratio, need not to be completely filled, non-additive or coagulation tube).

CLSI recommends that tubes are drawn in the following order, so that contamination by another anticoagulant agent than citrate or by a clot activator that may impact test result is avoided: (1) blood culture tube, (2) coagulation tube (light blue closure), (3) serum with or without clot activator, with or without gel (red closure), (4) heparin tube with or without gel separator (green closure), (5) EDTA tube with or without gel separator (lavender closure) and (5) glycolytic inhibitor (gray closure).

Influence of underfilling tubes:

The appropriate filling of coagulation tubes up to the nominal volume (as typically indicated on the tube) is essential, to produce the most appropriate blood-to-additive ratio, which is typically established at 1:10 (i.e., one part of anticoagulant plus nine parts of blood). Lippi et al. studied the influence of underfilling tubes on various coagulation assays and have identified a clinically significant bias in test results when tubes are drawn at less than 89% of total fill for activated partial thromboplastin time, less than 78% for fibrinogen, and less than 67% for coagulation factor VIII, whereas prothrombin time and activated protein C resistance remain relatively reliable even in tubes drawn at 67% of the nominal volume.

Effect of EDTA or heparin sodium versus citrate as the anticoagulant:

More and more plasma samples are referred to remote technical platfo resulrms. The nature of the anticoagulant of the anticoagulant present in the primary blood collection tube is not available in most of the cases. One frequent error is that EDTA plasma is referred instead of citrate plasma. This may result in erroneous results that are not necessarily easy to identify as results may mimic real clinical situations. PT and APTT are prolonged in EDTA versus citrate plasma samples whereas factor V and factor VIII levels are decreased; moreover, low factor V and factor VIII levels may make the laboratory testing for factor inhibitor, the result of which test will be falsely positive. Thrombophilia screening can also be affected in EDTA versus citrate plasma samples with decreased protein C and protein S activity, possible false positive lupus anticoagulant and possible no clot for activated protein C test.

Another possible error is testing on sodium heparin plasma aliquots instead of citrate plasma aliquots. Heparin samples are characterized by no clot PT, APTT and thrombin time, low factor VIII and factor IX levels, low von Willebrand ristocetin cofactor activity, low protein C and protein S activity (clotting assays) and high anti-Xa activity.

Wrong anticoagulant can easily be detected using routine chemistry tests: typical values for citrate plasma samples are 160mM for sodium, 3.2mM for potassium, 2.0mM for calcium and 0.79 for magnesium. EDTA plasma samples are characterized by a high potassium level (around 20mM), and very low or undetectable calcium and magnesium levels. Heparinized samples can be detected by high anti-Xa activity (roughly 3 IU/mL).

Whole blood sample transport:

Whole blood samples should be transported at room temperature. Transport on ice is not recommended as it may cause cold activation of factor VII, loss of von Willebrand factor, and platelet disruption. Extreme high and low temperatures should be avoided.

Ideally, whole blood samples should be transported to the lab within one hour of collection. However, PT has been demonstrated to be stable up to 24 hours, whereas the use of CTAD tubes can prevent platelet heparin neutralization up to four hours.

Special handling requirements should be provided to carriers.

Date, time shipped / received by the lab, and approximate temperature of the blood sample should be recorded upon arrival at the laboratory.

When using pneumatic tubes, specimens should be protected from vibrations and shock to avoid protein denaturation and platelet activation. Thalen et al. have studied the impact of pneumatic transport on platelet function and shown that the use of pneumatic tubes can affect platelet functions.

Precentrifugation sample examination:

Upon arrival at the laboratory, whole blood samples should be checked for gross clot formation by gentle inversion and observation by removing the cap and inserting and then removing two wooden sticks. However, (Micro)clots may not be detected by these methods. In a prospective study conducted on 1,334 samples, 0.6% (8) samples were found clotted. Differences observed between paired clotted and not clotted samples were not clinically relevant in routine tests (PT, APTT, fibrinogen) as well as for factor II and factor VII+X. In contrast, factor V levels were consistently higher in clotted versus not clotted samples.

Specimens that are clotted or collected in the wrong anticoagulant as well as collection containers that have other than the appropriate blood/anticoagulant ratio (under- / overfilled tubes) and mislabeled or unlabeled specimen should be rejected.


Centrifugation should be performed on capped specimen tubes for a time and at a speed that allow to produce platelet poor plasma (platelet count < 10 x 10/L or 10,000/µL). These conditions have to be established locally; this can be typically achieved by centrifuging whole blood samples at 1,500g for no less than 15 minutes. Higher speed and shorter duration (“Stat-fuge”) may be used in some instances (e. g. emergency situation).

Samples should be centrifuged at room temperature in order to avoid coagulation protein cold activation, with use of a swing-out bucket rotor to minimize contamination of plasma with platelets and other blood cells.

When critical, double centrifugation to ensure that the sample is platelet-poor (e.g. screening for lupus anticoagulant or before freezing plasma samples) is recommended.

The influence of centrifuge brake has been studied by Daves et al. who showed no statistical difference for APTT, a significant difference but with no clinical relevance (mean bias: 0.2 seconds) for PT and a statistical and clinically relevant although limited bias (mean bias: 0.29g/L) for fibrinogen.

Hemolysis, icterus and lipemia Interferences:

Hemolysis should be noted when visible. Lysis of red cells and resultant release of intracellular or membrane components may cause clotting factor activation that can have an impact on clotting time results whether using an optical or mechanical end-point detection system.

Icteric or lipemic samples or samples that contain substances that interfere with light transmission may generate erroneous results when using an optical end-point detection without accurate multi-wave length detection.

In a recent study, Wolley et al. studied the influence of hemolysis on PT run with two different reagents when using a viscosity-based dectection system (mechanical detection) in paired hemolyzed versus non-hemolyzed plasma samples. No significant difference was observed between hemolyzed versus non-hemolyzed samples.

In the same study, the impact of hemolysis for APTT when using three different APTT reagents was studied. No statistical difference was observed with one reagent, whereas a statistical difference was observed for the two other reagents, although the difference had no clinical impact for one these two reagents.

No statistical difference was observed for fibrinogen between hemolyzed and non-hemolyzed plasma samples.

Plasma sample storage:

Specimens should be stored capped unless sitting for brief periods (<30 minutes). Storage at refregirated temperature is not recommended: it may induce cold activation of factor VII, gradual loss of von Willebrand factor and factor VIII Plasma samples are usually stable for four hours at room temperature, two weeks frozen at -20°C and several months at -80°C. Stability should be checked locally for specific analytes.

For frozen sample storage, the use of frost-free freezers (automatic freeze/thaw cycles) is not recommended. Frozen samples should be rapidly thawed at 37°C, then gently mixed and tested immediately.

Importance of anticoagulant treatment information:

All anticoagulants may impact coagulation testing: vitamin K antagonists (e.g. warfarin, acenocoumarol, fluindione, phenprocoumon…), unfractionated heparin (UFH), low molecular weight heparins (e.g. enoxaparin, dalteparin, nadroparin, bemiparin, tinzaparin…), as well as the new direct oral anticoagulants, direct thrombin inhibitor dabigatran etexilate (PRADAXA®) and direct factor Xa inhibitors (rivaroxaban – XARELTO®, apixaban – ELIQUIS®, edoxaban – LIXIANA®, SAVAYSA®, betrixaban) can impact routine (PT, APTT, Thrombin Time – dabigatran) and specialized (factor assays, thrombophilia assays) clotting assays. Furthermore, even chromogenic assays if they use factor Xa or factor IIa (thrombin) in the assay principle may be impacted in presence of an anti-Xa or anti-IIa drug respectively.

Laboratories must know the assay principle of the test they use and the potential impact of anticoagulant treatment on these assays.

Information on patient’s anticoagulant treatment is crucial to avoid misinterpretation of test results.

Quality indicators in the preanalytic phase

Quality indicators are key for lab accreditation / certification. Quality indicators have been proposed as well as quality specifications in order to help clinical laboratories better manage their quality for preanalytics.

In conclusion, preanalytics is a multifaceted topic. Laboratories must be aware of the possible detrimental impact of preanalytical conditions on test results. Guidelines or similar documents are available from various organizations including the Clinical Laboratory Standards Institute (CLSI) that is quoted extensively in this paper, the International Council for Standardization in Haematology (ICSH), the College of American Pathologist (CAP), the Deutsches Institut für Normung (DIN), the Groupe Français d’Etudes sur l’Hémostase et la Thrombose (GFHT) or others. Clinical laboratories are highly encouraged to refer to them. They should, in the light of the relevant guideline(s), check their own procedures in order to minimize the risk of erroneous results caused by improper preanalytical conditions and monitor their performance using quality indicators.