4 Laboratory diagnosis
Recommendations
The global test activated partial thromboplastin time (APTT) will usually be prolonged and can be used as screening test in hemophilia A and B. The factor sensitivity for different APTT-reagents varies.
Factor VIII coagulation protein (FVIII:C) and FIX:C, functional activity assays can be measured with either one-stage clotting assay (OSA) or chromogenic substrate assay (CSA). Assay discrepancies can be caused by different mutations or treatments. OSAs are generally more prone to interferences.
Nijmegen-Bethesda assay is the recommended method for measurement of neutralizing anti-FVIII or anti-FIX antibodies. It is most reliable in patients without measurable factor activity (i.e. severe hemophilia).
Discrepant results between OSAs and CSAs (difference of 20-30%) exist, are reagent-dependent and are generally accepted when post infusion levels of recombinant extended half-life (EHL)-modified FVIII and FIX products are monitored.
For treatment with more than one modified EHL product, emicizumab or FVIII and FIX gene therapy, a difference larger than 30% has been obtained between different OSAs, or between OSA and CSAs.
For diagnostic and monitoring purposes, it is therefore of importance that the laboratory has access to more than one method for FVIII:C and FIX:C, respectively. Preferably one OSA and one CSA method.
4.1 Pre-analytical aspects of hemophilia testing
The pre-analytical phase is the time from patient identification and subsequent the blood collection until analysis of sample. In this phase, errors are often explained by incorrect specimen collection, handling, transportation, centrifugation or storage. Underfilling of tubes, or presence of clots due to incorrect mixing of tube, might lead to rejection of samples. The presence of anticoagulants in the sample, for example heparin contamination from a vascular access device, may interfere in the assay and give false test results. In order to reduce the pre-analytical error rate, it is important to understand the sources of variability and mechanisms that may lead to false assay results. Coagulation tests are exceptionally susceptible to suboptimal sample quality as the sample collection itself will initiate a hemostatic response. Thus, improper sample collection technique and/or incorrect sample handling prior to analysis will increase the risk of activation of the coagulation system in the tube, and incorrect test results may be obtained. In worst case incorrect results of screening and specific factor assays can lead to mismanagement of the patient. FVIII is one of the most labile coagulation factors and is degraded with time in vitro, and correct handling of the sample in the pre-analytical phase is very important. There are published guidelines/recommendations, how to assure sample integrity during the pre-analytical phase [15–17].
4.1.1 For plasma-based coagulation assays the recommendation is:
Direct venipuncture: Ensure atraumatic phlebotomy and minimal tourniquet.
Collection tube and order of draw: 3.2% (109 mmol/L sodium citrate, light blue stopper) first, or only after a non-additive tube.
Fill tube correctly to the mark (line), proportion blood to citrate 9:1.
Gently and thoroughly mix blood with anticoagulant (invert tube immediately 5-10 times).
Transport the whole-blood tube capped promptly at room temperature (RT).
Centrifuge within 1 hour after phlebotomy at minimum 1500 g, 15 min, RT, to obtain platelet poor plasma (PPP).
If analysis cannot be performed within four hours, the plasma should be transferred to another tube and frozen at -70\(^\circ\)C for later analysis.
4.2 Screening for hemophilia
Coagulation screening assays, such as (APTT) and prothrombin time (PT), are important for the initial laboratory evaluation of patients with bleeding disorders and are available in most hospital laboratories. If congenital or acquired hemophilia A or B are present, the APTT is often prolonged and the PT (INR) remains within reference ranges. Furthermore, a mixing procedure in which patient plasma is mixed with pooled normal plasma (PNP) (ratio 1:1) may distinguish between a coagulation factor deficiency and the presence of an inhibitor.In congenital hemophilia the APTT will be corrected after mixing of patient plasma with PNP [17]. If a mixing procedure does not correct the prolongation of APTT it may indicate the presence of an inhibitor (against a coagulation factor or an unspecific inhibitor such as lupus anticoagulant) or presence of interfering anticoagulants in the plasma.
There are several commercially available APTT reagents that vary in their sensitivity for coagulation factor deficiencies. To be accepted as a screening reagent to detect coagulation factor deficiency, it is recommended that the APTT reagent should give a prolonged clotting time at a factor activity of \(\leq\) 30% [17]. It is also important to mention that APTT is a global test that measures the collective activity of ten different coagulation factors. Certain conditions, for example an acute phase response leading to increased FVIII and fibrinogen, may shorten the APTT. Therefore, slightly reduced activity of one coagulation factor such as FVIII or FIX, compatible with mild hemophilia A or B, may have an APTT within reference ranges because one or several other coagulation factors are increased. Some APTT reagents are relatively insensitive to lupus anticoagulant and these may be advantageous. Thus, the treating physicians must be familiar with the characteristics of methods, and reference intervals.
4.3 Specific Factor VIII and IX assays
FVIII:C or FIX:C in plasma represents the functional (coagulation) activity of the factors and can be measured with either OSAs or CSAs [17,18]. These analyses are important in the diagnostic setting, during therapy (measurement of trough and peak levels after administration of replacement products) and also to detect the presence of inhibitory antibodies. When a family history is present, umbilical cord blood may be tested in male infants at birth to determine FVIII:C or FIX:C. For prenatal diagnosis, see chapter “Hemophilia in women and girls and hemophilia carriers”.
The FVIII:C and FIX:C assays should be calibrated with material that has traceability to current international standards for FVIII or FIX in plasma [17]. In this way the unit is given in international units (IU) and one IU is the factor activity present in one mL of normal plasma. In most of the Nordic countries the results are given in kIU/L or IU/mL but IU/dL is also used (1 IU/dL = 1%).
4.3.1 Differential diagnoses
Before ordering of samples information about case history, inheritance pattern, bleeding score and type of the bleeding are important to obtain. Once an increased APTT, in combination with a decreased FVIII activity, have been confirmed, there are several potential differential diagnoses such as congenital hemophilia A, acquired hemophilia A or von Willebrand disease (VWD), type 2N VWD (Normandy) in particular. A subsequent mixing procedure where APTT may correct or not, may help to distinguish between a coagulation factor deficiency and the presence of an inhibitor. Presence of unspecific inhibitors in plasma such as lupus anticoagulant, or contaminating anticoagulants, should be ruled out. In case of suspicion of a specific inhibitor, measurements of antibodies against FVIII and VWF activity are performed, and, when appropriate, the VWF:FVIII binding assay which determines the FVIII binding capacity of the patient’s VWF. Definite diagnosis may be dependent on sequencing of the F8 and VWD genes (see Genetic diagnosis paragraph).
4.3.2 Factor VIII:C assays
The OSA is the most frequently used assay principle in the world [17,18]. The main feature of the OSA is that it is based on the APTT test with the difference that the plasma is pre-diluted in FVIII-deficient plasma before analysis. In this way, the test system works with the simplicity of an APTT reaction but the pre-dilution procedure makes the FVIII activity in the sample the limiting factor, and thus determines the final clotting time. The ability of the sample to correct the APTT of a FVIII-deficient plasma can be expressed as FVIII:C activity if the assay is calibrated with a plasma with known concentrations of FVIII:C.
The performance of the OSA is affected by the FVIII-deficient plasma, the APTT reagent and the calibrator in use. The FVIII-deficient plasma can be obtained from a patient with severe hemophilia A (<0.01 kIU/L and without antibodies) or be immunodepleted. It is important to verify that new lots of FVIII-deficient plasmas are free from FVIII (<0.01 kIU/L) as this otherwise will compromise the test. Also, normal VWF concentrations of the FVIII-deficient plasma may be an advantage. The combination of reagents and instrumentation will also have an impact on the general assay characteristics. It is important that the laboratory methods can detect all hemophilia categories i.e. from mild to severe hemophilia A. The results obtained by OSAs may be affected by the presence of lupus anticoagulant, heparins, etc, and the influence is dependent on the APTT reagent.
CSA is another FVIII:C assay [17,18]. The assay procedure involves two separate reactions and the FVIII activity in the sample is the rate-limiting factor. There are several commercial CSAs available for measurements of FVIII:C in which the end product is color development from generation of activated FX that cleaves a chromogenic substrate. In the first step the diluted patient sample (or standard) is mixed with a reagent cocktail with purified factors IXa, X and phospholipids, leading to the formation of FXa. In the second step a specific chromogenic substrate for FXa is added. Cleavage of the substrate yield a color formation that is recorded spectrophotometrically, and the amount of color development is directly proportional to the FVIII:C activity in the sample. In general, the CSA has a lower detection limit than the APTT-based OSA, and due to the high dilution of the sample, the influence of interfering substances is less. CSA is commonly used among the Nordic hemophilia centers. CSA is also used by the pharmaceutical industry when the potencies of FVIII concentrates are assigned.
The two different FVIII:C assays give similar results in most cases. In general, the clinical phenotype corresponds better to the results of the CSA compared to the OSA. However, there is a significant assay discrepancy between OSAs and CSAs in approximately 20% of genetically confirmed mild/moderate hemophilia A patients, where FVIII activity measured by OSAs is at least 2-fold higher than FVIII:activity measured by CSAs. In such cases, i.e. patients with certain mutations in the F8 gene, mild hemophilia A may be missed if only a OSA is used [19]. However, there are also some genotypes causing mild hemophilia where there is an inverse assay discrepancy, and FVIII:C measured by OSAs may give the right diagnosis. Thus, to correctly identify mild hemophilia A may be challenging for the laboratory, if only one of the assay principles are used. For the management of hemophilia A, both a OSA and a CSA is recommended to be be performed to ensure detection of all new mild/moderate cases and to correctly assess the severity of the disease.
Reference interval: Usually between 0.50-1.50 kIU/L, but local differences may apply. Interpretation: Hemophilia A patients have low FVIII:C activity. FVIII:C activity <0.01 kIU/L are seen in severe hemophilia A. Moderate hemophilia A patients have FVIII:C activity between 0.01-0.05 kIU/L, and patients with mild haemophilia A have FVIII:C activity from >0.05 kIU/L up to 0.40 kIU/L. Carriers of hemophilia A have usually approximately 50% of the normal FVIII:C activity but occasionally they have similar FVIII:C activity as mild hemophiliacs.
FVIII is an acute phase reactant and the FVIII activity may increase several folds under certain conditions (e.g. trauma, infection, etc), and pregnancy/hormone treatments may also increase the FVIII activity. Mild hemophilia A, or a carrier state, can be missed if such conditions are not ruled out in the diagnostic setting.
4.3.3 Factor IX:C assays
The principle of OSAs for FIX is similar as for FVIII described above, with the only difference that the sample is prediluted in FIX-deficient plasma before analysis [20]. Thus, the main FIX:C assay principle is a test system based on an APTT reaction and the sample (patient or standard plasma) is pre-diluted in a plasma lacking FIX, which makes the activity of FIX in the sample the limiting factor. The OSA is calibrated with a standard that is traceable to the current international standard of FIX:C in plasma and results expressed as kIU/L or IU/dL (see FVIII:C above).
Measurement of FIX:C by CSAs has become commercially available, and is an alternative to the OSA. However, these CSAs for FIX:C are available at very few laboratories and have not yet been fully approved by regulatory bodies (EMA) for potency labelling. Nevertheless, local implementations in Nordic and other laboratories are encouraging, and it is possible that these CSAs will display analytical advantages compared to the OSAs as has been shown for measurements of FVIII:C. Furthermore, assay discrepancy, caused by certain mutations in the F9 gene, has also been described in hemophilia B [21].
Reference interval: Usually around 0.60-1.50 kIU/L but local differences may apply. Interpretation: Deficiency of FIX is the cause of hemophilia B. The degree of the deficiency is defined by the FIX activity. Severe hemophilia B patients have FIX:C <0.01 kIU/L, moderate hemophilia B patients have FIX:C activity between 0.01-0.05 kIU/L, and mild haemophilia B patients have FIX:C activity from >0.05 up to 0.40 kIU/L. Carriers of hemophilia B express about 50% of the expected normal FIX:C activity. Acquired hemophilia B, caused by specific inhibitors against the coagulation factor exists but is less frequent than the rare acquired hemophilia A.
4.3.4 Antibodies against FVIII or FIX
The development of neutralizing (inhibitors) and non-neutralizing antibodies (NNAs) is a complication to factor replacement therapy in hemophilia. The diagnostic methods generally available are based on defining the neutralizing capacity in functional assays, e.g. original or Nijmegen-modified Bethesda assays for antibodies generally considered to be inhibitors. Not only neutralizing antibodies can affect clearance of administered products though, therefore immunological assays can be used to define all antibodies, including NNAs. (In the Nordics, ELISA and xFLI methods are available in the Coagulation Lab in Malmö).
The hallmark of neutralizing anti-FVIII or anti-FIX antibodies (=inhibitors) is a prolonged APTT and normal PT/INR. The prolongation of the APTT is persistent also after mixing of the patient sample with an equal volume of PNP. Alloantibodies are most frequent and have a fast and dose-dependent antigen-antibody reaction. In acquired hemophilia A, time-dependent autoantibodies with a low-binding affinity can be present. For this reason, it is recommended to incubate the samples for two hours at 37\(^\circ\)C after a mixing procedure in order to allow the autoantibodies to have effect. Anti-FIX antibodies have faster kinetics and it is usually not necessary to incubate longer than 10 minutes. This incubation step is usually not necessary if screening for a FIX inhibitor.
In the Nijmegen-modified Bethesda assay, a test sample is prepared by mixing equal volumes of patient plasma with PNP and then measure the residual factor activity in the plasma mixture after two hours of incubation [17]. A control sample is prepared in parallel where PNP is mixed with an equal volume of FVIII-deficient plasma. Both test- and control samples are incubated for two hours at 37\(^\circ\)C and then the factor activities in both samples are measured. For anti-FVIII antibodies it is important to use buffered PNP (e.g. HEPES) as this stabilizes the pH and thus the FVIII activity during the incubation period, and reduces the risk of obtaining false positive results. There is usually a good correlation between the FVIII inhibitor results originating from the OSA and CSA FVIII:C assays [23]. Residual activity in the sample between 25 and 75% can be used for calculations of the inhibitor titer. By definition, one Bethesda unit (BU) is the inhibitor titer that neutralizes 50% of the factor activity in one mL plasma. If the residual activity is less than 25% it indicates an inhibitor titer above 2 BU/mL. Hence, these samples are prediluted in FVIII-deficient plasma before analysis. If several dilutions result in residual activities in the 25-75% range then the dilution that is closest to 50% is chosen for calculation of the inhibitor titer. It is recommended to perform the Bethesda assay when there has been a washout of the concentrate (i.e. FVIII:C activity <0.1 IU/ml) [23].
Reference interval: The cut-off for a positive result is by definition 0.4 BU/mL, as the recommendation was not to use any residual activity above 75% (75% residual activity corresponds to 0.4 BU/mL). To reduce the risk of false positive results many laboratories instead use 0.6 BU/mL as the cut-off for positivity because the test is less reliable in the low-titer range.
Interpretation: The presence of inhibitors may be suspected in patients with unexpected bleedings despite regular prophylaxis. This is also strengthened if the patient displays reduced recovery and half-life of the substituted factor. Patients with acquired hemophilia have very different clinical symptoms, caused by autoantibodies against factor VIII or IX.
Note: The Bethesda assay is usually performed on patients with severe type of hemophilia that do not have measurable FVIII:C (or FIX:C) activity. If the patient has an activity of 0.10 kIU/L or higher this must be taken into consideration when the inhibitor titer is calculated. It is recommended to remove the endogenous factor activity by heat-inactivation of the plasma sample at 56\(^\circ\)C before analysis [23].
4.3.5 Factor activity assays for monitoring treatment with EHL-products, emicizumab and gene therapy
For monitoring of postinfusion activity of full length recombinant factor VIII (rFVIII) standard products, CSA results are generally reported to be about 20% higher than OSA results. Well-documented discrepancies have also been described for the B-domain deleted rFVIII (Refacto®), with 30% higher results with CSAs compared to OSAs. Calibration with B-domain deleted FVIII has therefore been recommended for OSAs. For recombinant factor IX (rFIX) products, CSA results are instead in general 30% lower than OSA results [18].
Recently, a number of rFVIII and FIX products have been modified to obtain an EHL (i.e. pegylated, glycopegylated, and fusion proteins), and are available on the market or under development (see prophylaxis chapter) [18,24,25]. Some of the modifications affect the measurements of factor activity, and the effect is dependent on the method in use. As a consequence, both under- and overestimation of the factor activity occur in postinfusion patient samples, which might have a potential impact on patient management. This issue has been addressed by the European Medicines Agency (EMA). Ideally, the method used for potency labelling should be used for monitoring of factor activity in post-infusion samples. The European Pharmacopoeia (8th ed. 2016) recommends the use of CSAs for replacement factor potency labelling of FVIII and OSAs for FIX. The recommendation of the FVIII and FIX Scientific and Standardization Committee (SSC) of the International Society on Thrombosis and Haemostasis (ISTH) is that if either the OSA or the CSA for FVIII activity or both that provides provides valid potency estimates relative to the WHO IS for concentrates, they can be used for potency labelling. If there are discrepancies between assays, the most appropriate assay for labelling must be identified.
For several of the FVIII EHL-products, a difference of >30% in factor VIII activity has been obtained when different APTT-reagents have been used in different OSAs [24,25]. All reagent-product combinations have not been tested, and all reagents that contain the same activator (ellagic acid/phenol, silica/kaolin type) do not give the same results. More adequate and less variable factor VIII activity in postinfusion samples is generally obtained with CSAs. There are studies with comparable data for different modified products in the same test system [26,27]. ECAT and UK NEQAS (European external quality control providers in haemostasis) have performed larger field studies, and generalized product-by-product-based guidance about systematic under- or overestimation of factor activity for many reagent combinations, will be more available.
The use of product-specific reference standards was previously shown to reduce the discrepancy between OSAs and CSAs for B-domain deleted rFVIII, allowing a more accurate assessment of FVIII activity levels in patient plasma samples [18]. Another approach to avoid incorrect measurements of modified rFVIII and rFIX products in post-infusion plasma samples would be for the manufacturers to provide spiked samples for each laboratory to test before starting treatment monitoring of specific products. These approaches are however not practical options when many new products enter the market, and the laboratory must have information about the type of product present in every patient sample.
The factor VIII mimetic emicizumab and gene therapy for hemophilia A and B have become new treatment options that have required many haemostasis laboratories to evaluate new assays for monitoring [28]. Emicizumab affects APTT, which is shortened, and measurements of factor activity by APTT-based OSAs lead to falsely elevated factor activity, and these assays are not recommended to use.
CSA with bovine FX is insensitive to emicizumab, and in the presence of emicizumab, the endogenous activity of FVIII, the activity of rFVIII in post-infusion samples, and the concentration of neutralizing anti-FVIII antibodies must be measured with a CSA with bovine FX.
Emicizumab concentration can be measured with a modified (diluted) OSA or human CSA with emicizumab calibration. The results may be expressed either in emicizumab concentration or indirect FVIII activity. Anti-emicizumab antibodies may be indicated by a prolongation of the APTT.
There are issues with measurements of FVIII and FIX expressed from gene transfer, with approximately 1.5-fold higher FVIII activity variable depending on reagent and could be even higher for FIX activity measured by OSAs compared to CSAs [28]. The question under investigation is which assay system correlates with the hemostatic effect.
Other laboratory methods that are entering the clinical routine settings, are global tests, such as viscoelastic methods (ROTEM®, TEG®) and thrombin generation assays (CAT®, Ceveron alpha®), that can, if available, be used as a complement for monitoring certain products. It is of importance that each laboratory can validate the performance of methods used for monitoring treatment [27].
For the management of hemophilia patients with EHL-products, emicizumab and gene therapy, it is important that the laboratory has access to more than one method for FVIII and FIX respectively, preferably one OSA and one CSA method. With the available information of today it seems that the CSA, at least for the measurement of FVIII:C, is the more consistent laboratory method that can adequately measure the levels of most replacement products.
In conclusion:
In post-infusion samples, modified rFVIII and rFIX replacement products yield significant APTT-reagent dependent recovery after measurements with OSAs, whereas the recovery is more consistent with CSAs.
There is limited comparable data from field studies comparing several EHL-products in OSAs with various combinations of reagents, and CSAs, from which generalized product-by-product-based guidance about systematic under- or overestimation of activity can be given.
Ideally, similar recovery results should be obtained in post-infusion samples with the same method as used for potency labeling, and these data must be available from the manufacturers of the products.
The measurements of new treatment options like the FVIII mimetic emicizumab, and after FVIII and FIX gene therapy, are challenging for the laboratory. Further studies are needed to explore assay discrepancies after gene transfer with FVIII and FIX.
A significant challenge to the laboratories, and also to the clinicians, will be to communicate to the laboratory the specific treatment used by each patient.
4.4 Genetic diagnosis
Genotyping is clinically useful to predict the risk to develop inhibitor and for carrier- and prenatal diagnosis. For a detailed description of the clinical interpretation of genetic variants, we refer to the 2019 UK Haemophilia Centre Doctors’ Organisation Good Practice Paper) [29]. Depending on the experience and competence of the hemophilia team and the local organisation of genetic services, a clinical geneticist or counsellor can be part of the hemophilia care team.
Genetic diagnosis of severe hemophilia A starts with screening for the intron 22 inversion of the F8 gene which is caused by homologous recombination involving intron 22 and related sequences outside the F8 gene [30]. Approximately 40% of cases of severe hemophilia A is caused by intron 22 inversion. Similarly, an inversion involving intron 1 found in 1-2% of severe cases can also be screened for with a PCR technique. In the remaining cases of severe hemophilia A as well as moderate and mild cases, the whole F8 gene, 26 exons, must be sequenced usually through Sanger or next-generation sequencing (NGS) methodologies. Most patients have their own unique mutation and today a broad spectrum of more than 2000 variants are known causing hemophilia A. Variants such as nonsense and deletions, “null-mutations”, will obviously cause severe hemophilia since the DNA reading frame will be altered, mRNA aberrant and the FVIII protein will not be synthesized. A missense variant will usually produce a dysfunctional protein with reduced clotting activity but may also result in a ‘neutral/silent variant’ or a polymorphism. In such cases it is important to know if the same variant has been reported previously in patients with hemophilia, in databases such as the European EAHAD Coagulation Factor Variant Databases or the American CDC Hemophilia Mutation Project databases CHAMP/CHBM [31,32]. The clinical interpretation of a new or an unpublished genetic variant in the F8 or F9 genes, as well as other genes, should be based on guidelines published by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology ([33]); as pathogenic – likely pathogenic – variant of unknown significance (VUS) – likely benign - benign. One may also use in silico variant prediction programs to evaluate the deleterious effect of a variant. In a few percentages, disease causing variants/alterations will not be found despite sequencing of the whole gene, some of these cases having a more complex genetic background. The MLPA technique (Multiplex Ligation-dependent Probe Amplification) may show deletions or duplications not revealed on conventional Sanger sequencing.
In hemophilia B, the 8 exons of the F9 gene are sequenced and in almost all cases the disease causing variant will be found. Inversions are not present in the F9 gene but some patients have complete gene deletions, a strong predictor for development of inhibitors or anaphylactic reactions on FIX treatment.
Carrier diagnosis in sporadic case of hemophilia A or B, which encompasses around 50-60% of all newly diagnosed cases, may be a problem. In about 70-80% the mother of a sporadic case also carries the mutation and is thus a carrier. In the remaining 20-30% of cases a pathogenic variant is not found and these women may be either true non-carriers or being somatic and/or gonadal mosaics, i.e. it is not possible with conventional less sensitive techniques to conclude if she is a non-carrier or carrier. Mosaicism may cause a problem when genotyping mothers of a sporadic case of hemophilia A with conventional techniques but may be detected by more sensitive techniques such as ddPCR or NGS0 [34,35]. Studies indicate that, depending on the type of pathogenic variant, approximately 20% may be gonadal mosaics [36]. However, in hemophilia B this seems to be unusual [37].
Prenatal diagnosis (PND) can be achieved by chorionic villus sampling during the 11 to the 13th week of gestation and karyotype analysis can be performed with the aim to determine fetal sex and in male fetuses to diagnose the pathogenic variant within 2-3 working days. The reasons for PND may be to prevent the birth of an affected boy by termination of the pregnancy, to prepare the obstetrical procedures or, for the parents-to-be, to psychologically prepare having a child with hemophilia. Later in pregnancy amniotic fluid can be used as source of fetal DNA. Fetal sex determination can also be made by Y-chromosome analysis in blood from the pregnant woman very early in pregnancy and thus avoiding invasive diagnostic procedures in pregnancies with female fetus. Pre-implantation genetic diagnosis (PGD) enabling the implantation of female or unaffected male embryos has become possible [38,39]. PGD is a demanding procedure which however may be indicated in selected cases.