OUP user menu

Laboratory Assessment of Factor VIII Inhibitor Titer
The North American Specialized Coagulation Laboratory Association Experience

Ellinor I.B. Peerschke PhD, Donna D. Castellone MS, MT(ASCP)SH, Marlies Ledford-Kraemer MBA, MT(ASCP)SH, Elizabeth M. Van Cott MD, Piet Meijer PhD,
DOI: http://dx.doi.org/10.1309/AJCPMKP94CODILWS 552-558 First published online: 1 April 2009


Quantification of inhibitory antibodies against infused factor VIII (FVIII) has an important role in the management of patients with hemophilia A. This article summarizes results from the largest North American FVIII inhibitor proficiency testing challenge conducted to date. Test samples, 4 negative and 4 positive (1–3 Bethesda units [BU]/mL), were distributed by the ECAT Foundation in conjunction with the North American Specialized Coagulation Laboratory Association and analyzed by 38 to 42 laboratories in 2006 and 2007.

Whereas laboratories were able to distinguish between the absence and presence of low-titer FVIII inhibitors, the intralaboratory coefficient of variation was high (30%–42%) for inhibitor-positive samples, and the definition of lower detection limits of the assay was variable (0–1 BU/mL). Most laboratories performed the Bethesda assay with commercially supplied buffered normal pooled plasma in a 1:1 mix with patient plasma. These data provide information for the development of consensus guidelines to improve FVIII inhibitor quantification.

Key Words:
  • Coagulation
  • Factor VIII inhibitor titer
  • Hemophilia
  • Proficiency testing
  • North American Specialized Coagulation Laboratory Association
  • ECAT

Inhibitory antibodies to factor VIII (FVIII) present a major clinical challenge as a complication of hemophilia A and as acquired antibodies in patients without hemophilia.1,2 Despite improvements in hemophilia A management during the last decade with safer clotting factor products and prophylactic therapy, inhibitor development continues to be a major complication of therapy, occurring in 3% to 52% of patients with severe hemophilia A.1 In these patients, polyclonal alloantibodies neutralize FVIII infused from clotting factor products, leading to increased bleeding complications and difficult surgical management.

FVIII inhibitors can be identified in the laboratory by a variety of clot-based methods37 or enzyme linked immunosorbent assays (ELISAs).8,9 Worldwide, the Bethesda clot-based assay is the most commonly used laboratory test to quantify FVIII antibodies in plasma.6 The classic Bethesda method quantifies FVIII activity that remains after patient plasma (or increasing dilutions of patient plasma made using imidazole buffer to dilute the inhibitor) is mixed with an equal part of normal pooled plasma (NPP), which serves as the source for FVIII, and incubated for 2 hours (allowing time for inhibitor to neutralize FVIII) at 37°C. A control mixture that compensates for FVIII lability is prepared using NPP and buffer. The Nijmegen modification of this assay increased the specificity of low-titer FVIII inhibitor measurements by buffering the NPP, used in patient and control mixtures, to pH 7.4 with imidazole buffer and using FVIII-deficient plasma in the control mixture and for preparing patient dilutions.7 Both adjustments serve to maintain the pH of the reaction mixtures for the 2-hour incubation period and thereby stabilize FVIII in the NPP. The Bethesda assay and Nijmegen modification are outlined in Figure 1

Figure 1

Schematic representation of FVIII inhibitor assay using the classical Bethesda assay and the Nijmegen modification of the Bethesda assay. BU, Bethesda units; FVIII, factor VIII. Differences between the 2 methods are circled.

The immunologic response in patients with hemophilia A with inhibitors is generally categorized based on the Bethesda titer.10 High-titer inhibitors are defined as peak inhibitor titers of at least 5 Bethesda units (BU)/mL.1,2 Patients with inhibitor titers less than 5 BU/mL are considered to have a low immunologic response to FVIII replacement, whereas adult patients with inhibitor titers of 5 BU/mL or more and pediatric patients with titers of 10 BU/mL or more are considered high responders.11 Patients with inhibitors of more than 5 BU/mL are generally refractory to FVIII replacement therapy and require treatment with FVIII bypass agents or recombinant activated factor VII.12 Based on international consensus, patient management is influenced by inhibitor titers. Clinically relevant inhibitor development is defined as the occurrence of at least 2 positive inhibitor titers and decreased FVIII recovery (pharmacokinetics).1,2

Once an inhibitor develops, an immune tolerance regimen is undertaken to eradicate the inhibitors.11 Bethesda titers have been defined to guide immune tolerance therapy, predict outcome, and inform effectiveness of treatment. For example, immune tolerance is expected to be more successful in patients with inhibitor titers less than 10 BU/mL.13 Thus, despite the importance of beginning therapy soon after inhibitor confirmation, deferring immune tolerance therapy until the titer is below 10 BU/mL is often preferable.11 In addition, immune tolerance regimens are considered to be of limited efficacy in patients with titers of more than 200 BU/mL.14

Successful tolerance induction also has been defined by laboratory criteria.11 Thus, failure to demonstrate a progressive 20% reduction in inhibitor titers during each successive 6-month period of uninterrupted immune tolerance treatment, beginning 3 months after initiation, is considered a poor prognosis for inhibitor eradication.15 In contrast, successful tolerance induction may be assumed after 3 successive undetectable FVIII inhibitor titers.11 An international consensus defines undetectable inhibitor titers as 0.6 BU/mL or less with normalized FVIII pharmacokinetics.15

Because patient management is strongly influenced by the FVIII inhibitor titer, the present study was undertaken to investigate methods used by clinical laboratories in North America for FVIII inhibitor quantification and to examine the accuracy and between-laboratory precision of results. With data submitted from 38 to 42 independent special coagulation laboratories in 2006 and 2007, this represents the largest study of its kind conducted in North America, assessing performance with true-positive and true-negative samples on multiple challenges, and offers important insight into FVIII inhibitor assay sensitivity, accuracy, and variability.

Materials and Methods

The North American Specialized Coagulation Laboratory Association (NASCOLA) is a nonprofit organization that provides proficiency testing for North American laboratories performing diagnostic testing for bleeding or prothrombotic disorders and that creates a forum for critical evaluation of coagulation testing procedures, reagents, and instrumentation, which may aid in developing guidelines for appropriate use, performance, and interpretation of coagulation tests and results. In 2006, NASCOLA began distributing proficiency testing modules for FVIII inhibitor quantification during 2 of 4 annual proficiency testing challenges. Proficiency testing specimens consisted of lyophilized plasma samples distributed by the ECAT Foundation (Leiden, the Netherlands). Analysis of proficiency testing results was performed for 4 proficiency testing surveys, designated 2006–2, 2006–4, 2007–2, and 2007–4, each containing 2 specimens. The number of participating laboratories varied slightly for each survey with 40, 38, 40, and 42 laboratories submitting results for each of the 4 surveys, respectively. In addition to reporting results, laboratories also answered questions about test methods.


Assay Results

Table 1 summarizes FVIII inhibitor titers for 8 individual samples distributed during 4 separate proficiency testing events. Owing to the small number of laboratories performing methods other than the Bethesda assay, it was not possible to compare results by method. However, no one method or reagent seemed to perform differently from the others. Outlier results could not be attributed to reagent or methodological differences and did not occur consistently in the same laboratory during the 4 successive proficiency testing challenges.

Mean values for FVIII inhibitor titers reported by NASCOLA laboratories for all 8 test specimens agreed closely with the expected value. However, coefficients of variation (CVs) ranged from 30% to 42% for inhibitor-positive samples for inhibitor titers ranging between approximately 1 and 3 BU. Variability of results was greater with samples having no detectable inhibitor, reflecting statistical differences between small numbers. There were no changes in performance noted between 2006 and 2007.

Interpretation of positive Bethesda titers is summarized in Figure 2. The reported lower-limit sensitivities ranged from 0 to 1 BU. No information was available on how these sensitivities were established.

Figure 2

Reported lower sensitivity limits for local Bethesda assays.

View this table:
Table 1

Assay Characteristics

Assay characteristics reported by participating laboratories are summarized in Table 2. Whereas all laboratories provided analytic results, not all responded to the method questionnaire. The total number of respondents is, therefore, provided separately for each question. The results indicate that the majority of laboratories (>75%) quantified FVIII inhibitor titers using a Bethesda assay, with 20% or fewer of laboratories performing the Nijmegen modification. The apparent increase in laboratories performing the Nijmegen modification between 2006 and 2007 reflects a difference in the participant pool rather than a change in laboratory practices.

The majority of laboratories used NPP, which was obtained commercially, as the source for FVIII (Table 2). Approximately two thirds of the laboratories reported using NPP that was buffered. More than 90% performed a 1:1 mix of patient plasma (or patient plasma dilutions) and NPP to detect the presence of neutralizing FVIII antibodies. A variety of diluents were used for preparing patient plasma dilutions, with the most popular choices being imidazole buffer for the Bethesda assay and FVIII-deficient plasma for the Nijmegen modification, respectively. Whereas sources of FVIII-deficient plasma were more varied, 60% of laboratories obtained buffered NPP from 2 commercial sources Table 3.

Baseline FVIII levels recovered after incubation of the control mixture (NPP with buffer or FVIII-deficient plasma) ranged from 35% to 110%. Mean values are provided in Table 2. By accounting for the deterioration of FVIII in NPP during the incubation period, this value is used to correct the residual FVIII activity in patient plasma dilutions following incubation with NPP. The laboratories reporting residual FVIII levels of 100% to 110% may have been referring to the starting concentration of NPP. During the survey interval, 41% to 53% of laboratories indicated use of NPP that was assayed for FVIII using a standard that was calibrated against a World Health Organization (WHO) standard. The remaining percentage of laboratories indicated that the level of FVIII in their NPP was not traceable to a WHO standard. This may indicate that these laboratories did not know the level of FVIII in their NPP or that they measured it themselves but did not use a WHO standard.


The Bethesda assay is widely used to monitor the development and progression of FVIII inhibitors.16 Results are subject to a number of assay variables that impact reliability and clinical interpretation. For example, FVIII stability in NPP is compromised by pH shift and reduced protein concentration resulting from dilution. This may lead to spuriously positive Bethesda titers. This problem has been addressed by the Nijmegen modification,7 which uses NPP buffered to pH 7.4 with imidazole and substitutes FVIII-deficient plasma for buffer in the control mixture and, if high titers are present, in preparing patient plasma dilutions.

View this table:
Table 2
View this table:
Table 3

The Nijmegen method has not been widely adopted by NASCOLA laboratories, possibly owing to the expense of FVIII-deficient plasma relative to buffer. However, it is interesting to note that more than 60% of NASCOLA laboratories used a commercial, buffered NPP for the control mixture and for making the 1:1 patient mixtures. These laboratories are not performing a classic Bethesda assay but a hybrid of it and the Nijmegen method, suggesting that laboratories are modifying practices.

The presence of a lupus anticoagulant can result in a significant false-positive error rate.17 This was not addressed by the current FVIII inhibitor surveys. Future proficiency testing challenges to evaluate the effect of the lupus anticoagulant and other interferences would be important. In particular, interferences by the presence of heparin, FVIII bypass agents, and recombinant activated factor VII would be of interest.

Although the Bethesda assay has been used as the “gold standard” to quantify FVIII inhibitors, it is not able to detect noninhibitory antibodies, only roughly quantifies inhibitory antibodies found in patients without hemophilia, and cannot determine inhibitor isotypes.18 FVIII inhibitor ELISAs do not have these shortcomings and also show better sensitivity in detecting low-titer inhibitors. Whereas ELISAs are suitable for large-scale rapid screening to detect possible FVIII inhibitors, they cannot quantify an inhibitor, if present. Some authors question the sensitivity of the ELISA method, claiming that a negative ELISA result may need to be reevaluated by the Bethesda method.19 However, work by Lindgren and colleagues20 suggests that ELISA sensitivities vary owing to differing sources of the FVIII target on the microtiter plate, with von Willebrand factor presence adversely affecting sensitivity.

Reliable FVIII inhibitor assays are required for diagnosis and management of patients with hemophilia A and to facilitate interlaboratory comparison of inhibitor titers across studies performed to monitor the safety of FVIII replacement therapy. Results from the present study demonstrate remarkable uniformity in assay characteristics among NASCOLA laboratories. Despite this apparent uniformity, the interlaboratory CV for low-level FVIII inhibitor titers ranged between 30% and 42%. An even larger interlaboratory CV (56%) was reported for high FVIII inhibitor titers in a pilot study conducted in 2005 in which 32 NASCOLA laboratories evaluated a test sample with a FVIII inhibitor titer of 15 to 20 BU/mL. This level of imprecision makes it difficult to apply international consensus guidelines for patient evaluation and management based on FVIII inhibitor titers.

CVs reported for FVIII inhibitor titers are considerably larger than those reported for FVIII determinations. An evaluation of NASCOLA proficiency challenges in 2004 indicated a CV of 19% (n = 28) and 12% (n = 30) for specimens with target FVIII levels of 35% and 100%, respectively. In addition, CVs ranging from 7.5% to 20.4% were reported in recent proficiency challenges conducted by the College of American Pathologists.21 The larger CVs documented for FVIII inhibitors titers in the present study are not surprising, given the added complexities involved in the Bethesda assay. Because most patients with hemophilia A receive care in a single treatment center, it would be of interest to know the intralaboratory CV.

Although significant conformity in assay conditions was reported by NASCOLA laboratories, there was marked variation in the interpretation of a positive titer, ie, the lower-limit assay cutoff. Values from 0 to 1.0 BU/mL were reported. The Bethesda titer is calculated traditionally using values of residual FVIII activity between 25% and 75%.3,4,21 Residual activity levels of 75%, 50%, and 25% correspond to inhibitor titers of 0.4, 1.0, and 2.0 BU/mL, respectively. Thus, by definition, the sensitivity of the Bethesda assay is 0.4 BU/mL.

The lower limit of detection and the CV around this cutoff are of clinical importance for the identification of new patients with inhibitors. Each laboratory must standardize its own assay and determine what is to be considered an abnormal or positive result.11 Given the variability inherent in Bethesda assay results, FVIII recovery studies (pharmacokinetics) are considered more reliable11 and should be used to validate laboratory results.

The present study serves as a starting point for developing consensus guidelines for FVIII inhibitor quantification. Two important sources of variation exist in FVIII inhibitor methods that affect FVIII inactivation. Stabilization of pH at 7.4, through the use of buffered NPP for preparation of patient plasma/NPP 1:1 mixtures and the control mixture, leads to better test sensitivity by reducing the heightened inactivation of FVIII seen when using unbuffered NPP. In addition, the use of FVIII-deficient substrate plasma maintains a “like-to-like” comparison between patient plasma dilution mixtures and the control mixture (protein concentrations in both are similar). Variability in the technical preparation of patient plasma dilutions can also impact FVIII inhibitor methods by affecting FVIII lability. This was not addressed in these surveys but is of importance because preparation of patient plasma dilutions is not an automated process. Variables such as time, temperature, and type of dilution (serial vs independent) all potentially affect results.

The present study also identified assay variables that affect result interpretation. These include variability inherent in FVIII activity assays, definition of the FVIII inhibitor assay lower limit of detection, and the FVIII level in NPP. In addition, inconsistencies in BU conversion charts used in different laboratories have come to attention. The source of these inconsistencies is yet to be identified.

This study offers important insight into FVIII inhibitor assay sensitivity, accuracy, and variability across NASCOLA laboratories. It represents the largest study of its kind conducted in North America for examining FVIII inhibitor quantification using true-positive and true-negative samples in successive surveys.


Members of NASCOLA Proficiency Testing Committee included the following: Agnes Aysola, Department of Laboratory Medicine and Pathology, University of Minnesota, St Paul; Larry Brace, Department of Pathology, University of Illinois at Chicago; David Chance, St Louis University Hospital, St Louis, MO; Wayne L. Chandler, Department of Laboratory Medicine, University of Washington, Seattle; Jeffrey S. Dlott, Quest Diagnostics, Chantilly, VA; Charles Eby, Department of Pathology and Immunology, Washington University, St Louis, MO; Kenneth D. Friedman, Blood Center of Wisconsin, Milwaukee; John Heit, Department of Hematology Research, Mayo Clinic, Rochester, MN; Stephen Johnson, Hematology Laboratory, Tufts Medical Center, Boston, MA; Joan C. Mattson, East Lansing, MI; George M. Rodgers, Division of Hematology, University of Utah Medical Center, Salt Lake City; and Rita Selby, Sunnybrook and Women’s College Health Science Center, Toronto, Canada.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
View Abstract