Quality issues in laboratory haemostasis

Introduction to preanalytical variability

Bleeding and thrombotic disorders are major healthcare problems, worldwide [Lippi G, Favaloro EJ, Franchini M. Laboratory diagnostics and therapy in thrombosis and hemostasis: from bedside to bench to bedside. Semin Thromb Hemost 2009; 35: 3–8]. The approach to screening, diagnosis and monitoring of most disturbances of the haemostatic balance encompasses an appropriate and a discretionary use of laboratory diagnostics [Favaloro EJ, Lippi G, Franchini M. Laboratory diagnostics in thrombosis and hemostasis: the past, the present, and the future. Semin Thromb Hemost 2008; 34: 579–83]. Nevertheless, total quality in coagulation testing is an essential condition for producing clinically reliable and usable data. Several problems in the total testing process occur for events outside the direct control or jurisdiction of the laboratories performing the tests. Although there is perception that the preanalitical variability can adversely impact on results of first- and second-line coagulation testing, only a few studies have addressed this issue in detail. As such, the vast majority of problems, extraordinarily magnified over the traditional laboratory diagnostic for the peculiarity of the tests performed, have been related to inappropriateness test request, lack of knowledge of biological and circadian variability, collection of samples at the wrong time, misidentification, use of inappropriate procedures for sample collection (venous stasis, type of device, needle gauge), collection of unsuitable sample for quantity (incorrect blood to anticoagulant ratio) or quality (inappropriate sample matrix, haemolytic and lipaemic specimens, contamination from infusion routes), improper management of the sample after collection (inappropriate mixing, unrestrained procedures for transportation, spinning, storage, freezing and thawing) and other miscellanea variables [Lippi G, Franchini M, Montagnana M, Salvagno GL, Poli G, Guidi GC. Quality and reliability of routine coagulation testing: can we trust that sample? Blood Coagul Fibrinolysis 2006; 17: 513–9/Favaloro EJ. Pre-analytical variables in coagulation testing. Blood Coag Fibrinolysis 2007; 18: 86–89/Favaloro EJ, Lippi G, Adcock DM. Preanalytical and postanalytical variables: the leading causes of diagnostic error in hemostasis? Semin Thromb Hemost 2008; 34: 612–34] (Fig. 1). All these variables can seriously affect the integrity of the sample and impact on the reliability of the tests performed, thereby producing clinical and economical consequences that might adversely impact on laboratory organization and patient outcome. As such, unsuitable samples might account for up to 5.5% of all the specimens received for routine and specialized coagulation in a clinical laboratory. The highest frequency is traditionally observed for samples refereed from paediatric departments. The problems more frequently encountered have been identified with samples missing after a doctor’s order (49.3%), haemolysis (19.5%), clotting (14.2%) and inappropriate volume (13.7%). In particular, specimens missing seem to prevail in the intensive care unit, surgical and clinical departments, whereas clotted and haemolysed specimens are those most frequently recorded from paediatric and emergency departments respectively [Salvagno GL, Lippi G, Bassi A, Poli G, Guidi GC. Prevalence and type of pre-analytical problems for inpatients samples in coagulation laboratory. J Eval Clin Pract 2008; 14: 351–3].

Sample collection

An appropriate process of sample collection is essential for laboratory diagnostics, especially in those areas of testing where preanalytical variables might critically impact on the reliability of test result and patient outcome, such as haemostasis. Phlebotomy suffers from a high degree of preanalytical variability. Besides identification errors, which occur in almost every area of laboratory testing and are associated with the worst clinical outcomes (misdiagnosis and administration of inappropriate therapy) [Lippi G, Blanckaert N, Bonini P, Green S, Kitchen S, Palicka V, et al. Causes, consequences, detection, and prevention of identification errors in laboratory diagnostics. Clin Chem Lab Med 2009; 47: 143–53], most of the problems arise from cumbersome venipunctures, caused by shortages of skilled staff and overloaded phlebotomists. As such, several preanalytical variables in this process might led to unsuitable samples for testing (e.g., haemolytic, clotted, insufficient samples), and include blood collection from wrong sites (e.g., varicose veins, veins of arm or hand from the side of a mastectomy), using unsuitable disposals (i.e., infusive devices, cannulae, butterfly needle devices) [Lippi G, Guidi GC. Effect of specimen collection on routine coagulation assays and D-dimer measurement. Clin Chem 2004; 50: 2150–2], after a prolonged venous stasis caused by the tourniquet [Lippi G, Salvagno GL, Montagnana M, Guidi GC. Short-term venous stasis influences routine coagulation testing. Blood Coagul Fibrinolysis 2005; 16: 453–8] and by using an inappropriate procedure. The use of inappropriate container (i.e., a wrong blood tube) is another major source of concern for coagulation testing, along with inappropriate filling and mixing of the tubes. In particular, underfilling of blood tubes for the classical coagulation tests might significantly alter test results and, as such, the current H21-A5 Clinical Laboratory Standards Institute (CLSI) guideline recommends that coagulation samples should be discarded if the evacuated tube contains <90% of the expected fill volume [Adcock DM, Hoefner DM, Kottke-Marchant K, Marlar RA, Szamosi DI, Warunek DJ. Collection, Transport, and Processing of Blood Specimens for Testing Plasma-Based Coagulation Assays and Molecular Hemostasis Assays: Approved Guideline, 5th edn. Wayne, PA: Clinical Laboratory Standards Institute, CLSI document H21-A5; 2008]. At variance with the previous indications, there is clear evidence that no clinically meaningful differences in results of testing can be observed between the first and the second blood tubes collected sequentially, so that the necessity for drawing a discard tube before that used for coagulation testing is circumstantial at best. Upon collection, blood samples should be properly mixed (e.g., gently inverted four to six times) to allow effective mixing between blood and anticoagulants, and without producing haemolysis, clotting or platelet activation [Lippi G, Salvagno GL, Montagnana M, Guidi GC. Influence of primary sample mixing on routine coagulation testing. Blood Coagul Fibrinolysis 2007; 18: 709–11].

Sample handling, storage, transportation and preparation

To ensure quality of coagulation testing once a specimen has been collected properly, it is important to adhere to standard recommendations for processing, transportation and storage. Using plasma on spun-down cells at room temperature, add-on tests for routine coagulation testing can be performed within an 8-h period, obtaining results similar to what would be obtained from testing unstored specimens [Neofotistos D, Oropeza M, Ts’ao CH. Stability of plasma for add-on PT and APTT tests. Am J Clin Pathol 1998; 109:758–63], except for activated partial thromboplastin time (APTT) as measured on samples of patients receiving unfractionated heparin therapy [Adcock D, Kressin D, Marlar RA. The effect of time and temperature variables on routine coagulation tests. Blood Coagul Fibrinolysis 1998; 9: 463–70]. The growing trend towards consolidation of laboratory diagnostics in large, centralized facilities networked with peripheral phlebotomy services further enhances the implications of sample stability and transportation. This implies that whole blood samples rather than separated plasma samples may arrive to the central laboratory from varying distances, under variable storage and transportation modes. As such, we have recently established that a maximum of 6-h storage of uncentrifuged specimens at either room temperature or 4°C is allowed to maintain test results within the analytical quality specifications for desirable bias [Salvagno GL, Lippi G, Montagnana M, Franchini M, Poli G, Guidi GC. Influence of temperature and time before centrifugation of specimens for routine coagulation testing. Int J Lab Hematol 2009; 31: 462–7]. Sample transportation should also proceed as per current recommendations, that is non-refrigerated, at room (ambient) temperature, in the shortest possible time. For in-hospital transport, the novel generation of pneumatic tube systems is a valuable option, as it reduces turnaround times and labour, without introducing preanalytical errors for analysis of routine haematology, coagulation parameters and platelet function [Wallin O, Söderberg J, Grankvist K, Jonsson PA, Hultdin J. Preanalytical effects of pneumatic tube transport on routine haematology, coagulation parameters, platelet function and global coagulation. Clin Chem Lab Med 2008; 46: 1443–9].

Freeze-thawing of stored specimens can also produce the degradation of some labile factors, especially factor V (FV) and factor FVIII (FVIII). Accordingly, a low FVIII, for example, obtained from a referral laboratory might be simply artefactual because of inappropriate freeze-thaw cycles. When using frozen specimens, it might also be important to warm the sample to 37°C for at least 10 min before testing, to ensure reversal of any cryoprecipitate formed during freezing, especially when analysing FVIII and von Willebrand factor (VWF) [Favaloro EJ, Nair SC, Forsyth CJ. Collection and transport of samples for laboratory testing in von Willebrand’s disease (VWD): Time for a reappraisal? Thromb Haemost 2001; 86: 1589–90/Favaloro EJ, Soltani S, McDonald J. Potential laboratory misdiagnosis of haemophilia and von Willebrand disorder due to cold activation of blood samples for testing. Am J Clin Pathol 2004; 122: 686–92]. Inappropriate samples are also difficult to be detected when received as secondary plasma aliquots. As such, we have recently developed two simple, quick and inexpensive algorithms to help in the differential identification of citrated plasma vs. other samples with 100% sensitivity and specificity, should there be suspicion of inappropriate collection. The former algorithm is based on the sequential measurement of potassium, calcium and sodium, the latter on potassium and sodium [Lippi G, Salvagno GL, Adcock DM, Gelati M, Guidi GC, Favaloro EJ. Right or wrong sample received for coagulation testing? Tentative algorithms for detection of an incorrect type of sample. Int J Lab Hematol 2009 Feb 7. (Epub ahead of print)]. As regards centrifugation, the CLSI guideline contains specific recommendation to centrifuge capped specimens at 1500 g for no <15 min at room temperature [CLSI]. Basically, the centrifugation time is inversely associated with residual blood cell elements in plasma, especially platelets. Nevertheless, statistically significant variations from the reference 15-min centrifuge specimens were observed for fibrinogen in samples centrifuged for 5 min at most and for the APTT in samples centrifuged for 2 min at most. Meaningful biases related to the desirable bias were observed for fibrinogen in samples centrifuged for 2 min at most, and for the APTT in samples centrifuged for 1 min at most. According to these experimental conditions, a 5–10 min centrifuge time at 1500 g may be thereby still suitable for primary tubes collected for routine coagulation testing [Lippi G, Salvagno GL, Montagnana M, Manzato F, Guidi GC. Influence of the centrifuge time of primary plasma tubes on routine coagulation testing. Blood Coagul Fibrinolysis 2007; 18: 525–8]. Results of recent investigations also attest that whole blood specimen centrifugation at different temperatures than that currently recommended are not likely to generate significant analytical or clinical biases [Lippi G, Salvagno GL, Montagnana M, Poli G, Guidi GC. Influence of centrifuge temperature on routine coagulation testing. Clin Chem 2006; 52: 537–8]. The use of relative centrifugal forces >1500 g is not usually recommended, because they might induce platelet activation and haemolysis. For some tests, it might be preferred to use ‘double-spun’ plasma, to prevent inappropriate testing issues that may otherwise occur when using once-spun plasma. The use of filtered plasma is no longer recommended [Favaloro EJ, Lippi G, Adcock DM. Preanalytical and postanalytical variables: the leading causes of diagnostic error in hemostasis? Semin Thromb Hemost 2008; 34: 612–34].

Last but not least, it was recently reported that a plasma layer stratification might occur in primary tubes for coagulation testing, thereby introducing a substantial bias in measurements and yielding to shortened prothrombin time (PT) values (−1.0%) and higher fibrinogen concentrations (+1.2%), in the lower than in the upper part of a primary collection tube. As such, it has been suggested that plasma should be separated from the pellet immediately after centrifugation and appropriately mixed before delayed/repeated analysis or preparing aliquots [Lippi G, Salvagno GL, Bassi A, Montagnana M, Poli G, Guidi GC. Dishomogeneous separation of citrated plasma in primary collection tubes for routine coagulation testing. Blood Coagul Fibrinolysis 2008; 19: 330–2].

Survey on quality assurance practices in laboratories offering genetic tests for haemostasis disorders

A two-part survey was developed of which the first was designed to collect data about the laboratory setting, the director, supervisor and technician qualifications and their experience, the type and number of testing services provided for genetic diagnosis of haemophilia A and haemophilia B, thrombophilia and rare bleeding disorders. The second-part of the survey included questions on laboratory policies and practices in place during the pre-analytical, analytical and postanalytical stages of genetic testing. It also included information on policies regarding informed consent, the transportation of specimens, samples and genetic data handling, reporting practices and participation in proficiency testing programmes. Of the 98 potential laboratory directors contacted based on information gathered from the published literature and from the World Federation of Haemophilia training centre directory, we have received a response from 17 laboratories so-far [April 2010]. These included laboratories from Argentina, Australia, Belgium, Brazil, Canada, China, Germany, India, Italy, Japan, Netherlands, Thailand and UK. Table 1 summarizes the data collected from the first part of the survey. Laboratories from developed countries (n = 10) had more experienced personnel (21.5 years vs. 12 years) performing genetic diagnoses compared with those from developing countries (n = 7) possibly because such genetic tests are available only from the last decade in developing countries. Most of the laboratories (94%) surveyed seem to be in close proximity to, or developed within clinical centres to provide comprehensive patient care. This also explains a high frequency (∼60%) of integrated diagnostic service provided by these laboratories for three of the following four conditions (haemophilia A, haemophilia B, rare bleeding disorders and thrombophilia). A majority of the laboratories surveyed (88%) also accept DNA samples from other centres implying cross-country or trans-border referral has become common as centre’s specializing in rare genetic disorders are relatively few, especially in developing countries. Among the strategies used for molecular genetic testing of haemophilia (Table 1), there was no perceptible geographical disparity suggesting that advanced molecular techniques have successfully been incorporated in the testing algorithms of the laboratories in developing countries also. However, a potential bias cannot be excluded, as the responses from developing countries have been largely from the centres that have been well recognized for their research in bleeding disorders. It is possible that smaller laboratories performing techniques such as linkage analysis have been under-represented in their response to this survey.

Examination of the quality assurance practices in the surveyed laboratories showed wide variability. Although the majority of laboratories (75%) had separate rooms for DNA extraction and PCR, this was not universal. Access to the laboratory was controlled and diagnostic laboratory data were stored in a secure area in 75% of laboratories. Validated reagents were used during genetic testing in 69% of laboratories, or had separate pipettes for aliquoting primers and DNA or for PCR in 88%. Periodic calibration of equipments used in genetic testing such as the DNA sequencer was performed in 88% of laboratories. A similar number required an informed consent for genetic diagnosis (88%). Reporting practices on the molecular diagnosis were also not uniform. Only 69% of laboratories provided written reports to patients, and only 60% of the reports had a description of clinical evidence examined. Sub-optimal quality assurance practices were more prevalent in laboratories from developing countries. This is further reflected in the fact that only 43% (vs. 100% in developed countries) of laboratories in developing countries participate in any kind of proficiency testing programmes.

Establishing an external quality assurance scheme for genetic testing of haemostasis disorders in India

External quality assurance scheme (EQAS) for genetic tests are rarely available in developing countries. To address this, we have initiated an EQAS for molecular genetic analysis of haematological disorders for laboratories in India, since 2006. Two surveys are conducted each year. Nine laboratories currently participate in the programme. Previously characterized DNA samples, obtained with informed consent, from patients with haemophilia A, haemophilia B and thrombotic disorders are sent for analysis along with brief clinical details. Six EQA cycles have been completed. For haemophilia A, linkage analysis was most commonly used with only one laboratory performing direct mutation analysis using conformation sensitive gel electrophoresis (CSGE) and DNA sequencing. Assessment of haemophilia B was performed by only ∼40% of laboratories and is mostly performed by CSGE. All laboratories involved in thrombophilia testing use PCR-RFLP for genotyping. Response rate for this EQAS schemes are about 70% with a reporting accuracy of over 90%. A performance report is provided to all participants. Our experience suggests that EQAS for genetic tests can be effectively established in developing countries.

In conclusion, our survey shows that quality assurance practices have not penetrated diagnostic molecular genetic testing laboratories consistently across the world and in developing countries in particular. This underscores the need for increasing the awareness and benefits of voluntary participation in laboratory accreditation programmes, establishing EQA programmes that are cost-effective, and formulate specific guidelines for standardization of proficiency testing. To increase participation in developing countries, participation in such programmes should be made mandatory.