THE DIAGNOSIS and MANAGEMENT OF FACTOR VIII and IX INHIBITORS: A GUIDELINE FROM THE UK HAEMOPHILIA CENTRE DOCTORS' ORGANIZATION (UKHCDO)

The expected incidence of inhibitors in severe haemophilia A and B approaches 33% and 3% respectively (Katz, 1996). The occurrence of this complication has significant clinical implications, as the response to treatment becomes uncertain, morbidity is increased and life expectancy reduced. Direct medical costs are much higher for inhibitor patients (Goudemand, 1998), as are direct non-medical costs to the patient, their family and to society as a whole. This must be fully recognized by commissioners or purchasers of haemophilia services. Since the previous guideline on the detection and management of factor VIII inhibitors was published, significant diagnostic and therapeutic advances have taken place (Hay et al, 1996a). The UK Haemophilia Centre Doctors' Organization (UKHCDO) has therefore revised, updated and substantially rewritten the earlier guideline. In doing so, we have tried to define best current practice internationally, as reflected by the literature. We have avoided recommending one haemostatic product over another where no direct comparative trials have been conducted. Where marked national differences in clinical practice exist, in immune tolerance induction for example, we recommend the strategy for which the highest level of evidence exists. This evidence-based approach highlights the need for future clinical trials in areas where current treatment strategies are based on uncontrolled observations and where there is a dichotomy of clinical opinion. The guidelines were drafted by the UKHCDO Inhibitor Working Party and circulated to the Executive Committee of the UKHCDO for consultation and editorial commentary. These include several members who practice only paediatric haematology and two members who are primarily general haematologists. Members of the working party make an annual declaration of interest to UKHCDO. Relevant scientific papers were identified from Medline using the index terms h(a)emophilia, factor VIII and IX, inhibitors, antibodies, alloantibodies, autoantibodies and management [Agency of Health Care Policy and Research (AHCPR), 1992]. Recommendations were based on reports with the highest levels of evidence available (see Appendix). Inhibitors develop in patients with severe haemophilia A after a median of 9–12 treatment days (Ehrenforth et al, 1992; Addiego et al, 1993; Lusher et al, 1993). Outbreaks of inhibitors among frequently treated adults normally considered to be at low risk of inhibitor development have also been reported in relation to changes in factor VIII concentrate manufacture (Peerlinck et al, 1993; Rosendaal et al, 1993; Rosendaal (1997). Therefore, patients should be screened more frequently during the early phase of treatment and following any change to a new and un-established type of factor VIII concentrate. Patients with factor IX deficiency should be monitored in the same way. In vitro laboratory tests for the detection and quantification of factor VIII antibodies will detect those neutralizing inhibitor antibodies that interfere with the function of factor VIII or IX. Factor VIII recovery and plasma half-life studies are required for the detection of inhibitor antibodies that reduce factor VIII survival and for the detection of very low-level inhibitors. Children on regular prophylaxis and lacking free antibody are an example of this. Patients should be screened for inhibitors after every 5th exposure day or every 3 months until the 20th exposure day, every 6–12 months thereafter and prior to any surgical procedure. Patients should also be screened for inhibitors if the frequency of bleeding increases or if the clinical or laboratory response to replacement therapy is poor (grade B recommendation based on level III evidence). If the clinical response or factor VIII increment are poor, an estimate of factor VIII recovery and/or half-life should be conducted, particularly if the inhibitor assay is negative (grade B recommendation based on level III evidence). Screening for inhibitors is normally conducted using an activated partial thromboplastin time (APTT)-based or Bethesda method, but children on regular prophylaxis should be screened using factor VIII/IX recovery or half-life measurements (commonly peak and trough factor VIII/IX measurements in the first instance) (grade C recommendation based on level IV evidence). Factor VIII inactivation by inhibitors is time and temperature dependent (Lossing et al, 1977) and so the APTT of the patient:pooled-plasma mixture should be measured immediately after mixing and after incubation. The most widely used APTT screening method for factor VIII inhibitors compares the APTT of a patient:pooled-plasma mixture immediately after mixing and after 2 h incubation (Ewing & Kasper, 1982). Each laboratory must standardize this test independently and determine what they consider an abnormal result. It is recommended that an APTT-based method should be used to screen patients for factor VIII inhibitors (level 1b recommendation based on grade A evidence). The Bethesda assay has been recommended as the standard method for measuring the factor VIII inhibitor titre (Kasper et al, 1975). The original Bethesda method may give false-positive results owing to loss of factor VIII activity caused by a pH shift and reduced protein concentration unrelated to the presence of an inhibitor. The Nijmegen modification of the assay avoids this pH shift (Verbruggen et al, 1995). False-positive results are eliminated by this modification, which has now been recommended by the International Society of Thrombosis and Haemostasis Factor VIII/IX Scientific Sub-committee (Giles et al, 1998). The Bethesda assay can be modified to measure the degree to which an inhibitor inactivates porcine factor VIII using porcine factor VIII concentrate as substrate (Hyate:C, Ipsen, UK). This is diluted in haemophilic plasma to a concentration of 100 IU/dl and measured in a standard factor VIII assay using the usual reagents and standards. Factor IX inhibitory activity may be quantified using the Bethesda method, but omitting the two-hour incubation (Ewing & Kasper, 1982). There are no published data on the role of the Nijmegen modification for the quantification of factor IX inhibitors. Factor VIII inhibitors should be quantified using the Nijmegen modification of the Bethesda assay (grade A recommendation based on level Ib evidence). Factor VIII inhibitor titre should initially be quantified to both human and porcine factor VIII and, subsequently, if treatment with porcine factor VIII is contemplated (grade B recommendation based on level III evidence.) Factor IX inhibitors should be quantified using the Bethesda assay without an incubation period (grade B recommendation based on level III evidence). The diagnosis is established by the demonstration of an isolated, time-dependent, prolongation of the APTT. Specific tests for the lupus anticoagulant should be negative. A marked reduction in factor VIII is commonly accompanied by more modest in vitro reductions in factor IX, XI and XII activity, which may create the false impression that the inhibitor is non-specific. This is a laboratory artefact caused by depletion of factor VIII in the substrate plasma, which may be 'diluted out' so that assays of increasing dilutions of the patient's plasma give progressively increasing factor IX, XI and XII levels while having no effect on the uniformly low factor VIII level. The inhibitor titre should be determined using both human and porcine factor VIII because factor VIII auto-antibodies usually exhibit little inhibitor activity against porcine factor VIII (Kasper, 1991; Fiks-Sigaud et al, 1993; Morrison et al, 1993). The Bethesda assay may underestimate inhibitor potency in acquired haemophilia because the second-order reaction kinetics result in the persistence of low levels of factor VIII, even after prolonged incubation. It has been suggested that, for consistency, in this situation, one should report the inhibitor titre calculated from the lowest dilution that results in approximately 50% residual factor VIII after the 2 h incubation (Kasper, 1991). The diagnosis of acquired haemophilia should be based upon the clinical presentation, an isolated prolongation of the APTT that corrects initially with normal plasma, but which prolongs on incubation, a marked reduction in factor VIII concentration (usually < 0·01 IU/ml) and an inhibitor measurable using the Bethesda Assay. Specific tests for the lupus anticoagulant should also be negative. It is also recommended that the inhibitor in acquired haemophilia should be quantified using both human and porcine factor VIII, calculated from the lowest dilution that results in approximately 50% residual factor VIII (grade B recommendation based on level III evidence). For pharmacokinetic studies in adult patients, samples are taken at several time-points up to 2 h after infusion to establish the peak factor VIII value. Recovery is calculated from the peak plasma factor VIII concentration using a formula that requires an estimation of plasma volume and which is unsuitable for paediatric use (Kjellman, 1984; Kasper, 1991; Morfini et al, 1991). In routine clinical practice, a single measurement taken 30–60 min after infusion is usually used. Factor recovery percentage should be calculated with reference to the recovery constant (k) for that product, as this constant may vary from one product to another. k is commonly taken as 2·0 for factor VIII and 1·0 for factor IX, but different values apply to some factor VIII concentrates and to recombinant IX (Benefix, Genetics Institute, USA). Normal recovery values range from 75% to 100% (Kjellman, 1984). Recovery and half-life are commonly lower in children. A factor VIII recovery as low as 66% and a half-life as low as 6 h may be observed in small haemophilic children lacking inhibitors. There are minor differences when recovery is calculated using one- or two-stage assays and recovery may be 20–30% higher when the chromogenic assay method is used (Lee et al, 1996). Pharmacokinetic studies of B-domain-deleted recombinant factor VIII (BDDrFVIII, Refacto, Wyeth, USA) should be conducted using either the chromogenic method or a one-stage method that uses a specific standard or a thromboplastin optimized for this product. The standard one-stage factor VIII assay may underestimate BDDrFVIII recovery by 30–50%. Factor VIII recovery should be based upon the difference in factor VIII concentration between a sample taken pre- and 30–60 min after infusion (grade C recommendation based on level IV evidence). B-domainless factor VIII should be measured using either a chromogenic assay or a one-stage assay using a specific, standard (grade B recommendation based upon level IIb evidence). Factor VIII recovery in children should be calculated using a simple formula that requires no estimate of plasma volume (grade B recommendation based upon level III evidence). Factor VIII activity-time curves fit a biphasic two-compartment model. The initial decline in factor VIII is the distribution (disappearance) half-time and is complete within 4 h. The second slope of decay is the elimination half-life. The half-disappearance time is the time from infusion until the factor VIII concentration has fallen to 50% of the peak post-infusion level. Short half-life studies are influenced more by the distribution half-life than by the elimination half-life and so it is recommended that half-life studies are continued for at least 24–48 h after infusion. Model-dependent or model-independent analysis may be used to analyse the data, but will not give the same results (Lee et al, 1990; Morfini et al, 1991). Different computer programmes used to analyse the same data set may also give different estimates of pharmacokinetic parameters. Although a non-compartmental model is less subject to this variability, it may give different results from a two-compartment model (Pascual & Montoro, 1997). In most half-life studies, 50 U/kg of factor VIII or 75 IU/kg of factor IX are infused after a washout period of at least 72 h or when the baseline factor level is reached (typically < 0·01 u/dl). Samples should be taken regularly for 48 h or until the factor activity has fallen to baseline levels. In the absence of a pharmacokinetic computer model, the elimination half-time should be calculated by linear least-squares regression analysis (Kjellman, 1984; Kasper, 1991). At least four sample-points are required to establish a log-linear phase (Kasper, 1991). Half-life studies should be conducted after a wash-out period or when the factor VIII or IX level has reached baseline. A sample should be taken at 1 h, 4 h and at several points thereafter until the factor VIII activity has fallen to baseline (grade B recommendation based on level III evidence). The two aspects of inhibitor management, inhibitor abolition through immune tolerance induction and the haemostatic management of bleeding episodes and surgery, will be reviewed separately. Immune tolerance induction (ITI) requires study on an international collaborative basis if we are to learn how this approach may be optimized and applied in the most cost-efficient way. ITI must be viewed as a long-term investment and the period of intensive treatment compared with the cost of life-long treatment in the presence of a persistent high inhibitor titre. Treatment of acute bleeding must be active and initiated early, as this aggressive approach should reduce patient morbidity or mortality and also reduce overall direct medical costs. It is important that commissioners and treaters enter into partnership so that the clinical effectiveness, cost benefit and cost-effectiveness of immune tolerance and other treatment approaches are studied and analysed prospectively by national and international studies. The management of such patients should be supervised by a Haemophilia Comprehensive Care Centre, as defined in the NHS Executive Health Service Guidelines (HSG) 30 (1993). Factor VIII inhibitors may be abolished in more than 80% of selected patients with severe haemophilia A ITI (Nilsson et al, 1988; Mariani et al, 1994; Kreuz et al, 1995; Mauser-Bunschoten et al, 1995; Brackmann, 1996; DiMichele et al, 1999,2001). Successful ITI leads to normalization of the factor VIII half-life, a marked improvement in the patient's quality of life and a considerable reduction in the future cost of treatment. Current knowledge of ITI for factor VIII and IX inhibitors is derived from uncontrolled series of patients treated using various factor VIII dose regimens and the results of three retrospective surveys of ITI. These surveys include the International Immune Tolerance Registry (IITR, Mariani et al, 1994), the North American Immune Tolerance Registry (NAITR, DiMichele et al, 1999, 2001) and the German Immune Tolerance Registry (GITR, Lenk, 1999). There are no controlled comparisons of the regimens currently used for ITI and no agreement on the optimal regimen to be used. The most important predictor of successful ITI is the inhibitor titre at the start of ITI, which affects both the likelihood of success and the time taken to achieve tolerance. An inhibitor titre of < 10 Bethesda units (BU)/ml at the time of initiation of ITI significantly correlated with successful outcome in both the NAITR and the IITR (P = 0·004 and 0·001 respectively) (Mariani et al, 1994; DiMichele et al, 1999,2001; Lenk, 1999). The success rate and time to success for patients starting ITI with an inhibitor titre of < 10 BU/ml was 85% and 11 months compared with 43% and 15 months for patients with inhibitors of > 10 BU. Most other studies show a similar relationship between the starting inhibitor titre, the outcome and the time taken to achieve tolerance (Kreuz et al, 1995; Mauser-Bunschoten et al, 1995). A low peak historical inhibitor titre prior to ITI has been said to predict successful ITI, but this variable was far less strongly related to outcome than the inhibitor titre at the start of ITI in the IITR or NAITR (Mariani et al, 1994; DiMichele et al, 1999). Very high starting inhibitor titres of > 500 BU/ml are associated with resistance and a poor response to ITI (level IIb, Mariani et al, 1994; Kreuz et al, 1995; Mauser-Bunschoten et al, 1995; DiMichele et al, 1999). Although it is widely believed that ITI should start as soon as possible after the inhibitor is detected, there is no firm scientific basis for this approach. A short interval between inhibitor detection and the initiation of ITI predicted a successful outcome in some studies (Mariani et al, 1994; Kreuz et al, 1995), but not others (Mauser-Bunschoten et al, 1995; DiMichele et al, 1999, 2001). The chance of achieving successful ITI should be enhanced by deliberately deferring the initiation of ITI until the inhibitor titre has declined below 10 BU/ml and preferably below 5 BU/ml, as the success of ITI relates significantly to the starting inhibitor, but not to the peak historical inhibitor titre. Series in which ITI was deferred either deliberately or by circumstance until the inhibitor titre was < 10 BU/ml have been notably successful (Mauser-Bunschoten et al, 1995; Smith et al, 1999; Rocino et al 2000). These authors report similar success-rates of 88–100%, despite using widely varying factor VIII dose rates. The influence of the dose of factor VIII used is disputed. The IITR suggested that larger doses are significantly more effective, particularly in patients with inhibitor titres of > 10 BU/ml (Mariani et al, 1994). In contrast, neither the NAITR nor the GITR were able to demonstrate such a dose relationship (DiMichele et al, 1999, 2001; Lenk, 1999). Furthermore, the low-dose regime has been reported to achieve a success rate of 88% amongst a cohort in whom the inhibitor titre had declined to < 10 BU/ml before the initiation of ITI (Kreutz et al, 1995; Mauser-Bunschoten et al, 1995; Brackmann, 1996; Lenk, 1999). Tolerance may be induced more easily in younger patients whose inhibitors are not long established (Mariani et al, 1994; Kreuz et al, 1995; Mauser-Bunschoten et al, 1995), although this is disputed by DiMichele et al (1999,2001). There are no convincing data to suggest that any particular type or brand of factor VIII concentrate is more or less effective for ITI. Although Kreuz has suggested that patients may be more readily tolerized using intermediate-purity factor VIII concentrate, this data is inconclusive and is based on uncontrolled observations in six patients (Kreuz et al, 1996). Others have demonstrated similar success using high-purity or recombinant factor VIII concentrates (Smith et al, 1999; Rocino et al 2000). Many low-level inhibitors will disappear spontaneously without ITI, although troublesome inhibitors may also present with a low titre. It would be reasonable therefore for inhibitors presenting with a titre of 2 BU/ml or less to be monitored weekly for evidence of an increase in titre and to defer initiation of ITI. Tolerance is achieved by the regular administration of factor VIII or IX over a period of a few months to two or more years. Widely differing doses of factor VIII have been used, varying from 50 IU/kg three times a week to 300 IU/kg/d. Intermediate doses of 50 or 100 IU/kg/d are also widely used with success. The regimens in common use are summarized in Table I. Regimens that combine intensive factor VIII/IX replacement with concomitant immunosuppression have also been described and are outlined in the table (Nilsson et al, 1988). The best described of these is the Malmo regimen, in which high-dose factor VIII or IX replacement is combined with cyclophosphamide, high-dose immunoglobulin and protein A immuno-adsorption (Nilsson et al, 1988; & 1996). This regimen is not in use as immuno-adsorption is in small children and are also to use in this The approach to induction has not been Although high-dose regimens may achieve tolerance more it is not their overall success rate is to that using a low-dose regimens may be more readily than high-dose regimens without the use of and may also be more for the patient and factor VIII or IX replacement therapy for ITI may and the of therapy such as or or recombinant of ITI and should be avoided during the of ITI, as they may influence both success and the time taken to achieve tolerance (Kreuz et al, 1995; Brackmann, 1996; Lenk, 1999). immune tolerance the inhibitor should be quantified at regular until free inhibitor is no using the Bethesda Factor VIII recovery should be at until normal recovery is the factor VIII half-life should be determined at until it is also normal 6 Tolerance is taken as of normal factor VIII recovery and half-life. In North ITI is and factor VIII prophylaxis as soon as tolerance has been demonstrated (DiMichele et al, 1999). In it is more usual to ITI for several months after tolerance is established and to the factor VIII dose over 3 months before starting normal prophylaxis (Mariani et al, 1994; Kreuz et al, 1995; Brackmann, 1996; Lenk, 1999). This is not of that the rate of is very low of the patient's factor VIII dose is or There are few published reports of induction in haemophilia B as factor IX inhibitors are the regimens described have been used for haemophilia B with some success using doses of factor IX similar to the doses of factor VIII used for ITI of factor VIII inhibitors (Nilsson et al, 1988; DiMichele et al, 1999). to ITI in haemophilia B include the risk of the and a poor overall response rate to ITI. Immune tolerance should be using high factor IX concentrates or recombinant factor IX to the associated with high doses of concentrates patients with a of have ITI with factor IX, but most continued to with and et al, 1998). An between to factor IX and has also been reported in patients treated with doses of factor IX for ITI. The after a median of months ITI et al, et al, 1998). These patients not to but some following a dose reduction or of ITI et al, 1998). The and of ITI should be considered in patients with factor IX inhibitors and a of in of the low success rate reported and the high risk of the Immune tolerance induction is recommended for patients with haemophilia A or B and a factor VIII or IX inhibitor and should be considered as early as possible after the presence of an inhibitor has been (grade B level of evidence ITI should be conducted the of a Haemophilia Comprehensive Care Centre as defined by NHS Executive Health Service Guidelines (HSG) 30 (grade C recommendation based on level IV evidence). Immune tolerance induction is for both patients and and should be from the or before starting (grade C recommendation based on level IV evidence). The haemophilia the ITI should have of factor VIII/IX therapy or recombinant factor and for the of for IV (grade B recommendation based on level IIb evidence). It is recommended that, prior to the initiation of ITI, bleeding should be on with preferably using recombinant factor to an in inhibitor titre. ITI should be deferred until the inhibitor titre has fallen below 10 BU/ml preferably below 5 It is recommended that patients ITI be into comparative clinical trials of ITI for which they are or that data from their ITI be in one of the international of ITI (grade B recommendation based on level III evidence). Current UKHCDO is that immune tolerance in children the of should be conducted using recombinant factor VIII or IX, where in with the Health Service and current UKHCDO therapeutic Guidelines (grade B recommendation based on level evidence). of ITI should be avoided (grade C recommendation based on level IV evidence). immune the inhibitor titre should be until free inhibitor is no Recovery should be determined until normal and half-life determined every 3 months until tolerance is Tolerance is defined as the of normal factor VIII recovery and half-life (grade C recommendation based on level IV evidence). tolerance has been achieved it is recommended that factor VIII or IX prophylaxis start without ITI or of the factor VIII/IX dose (grade B recommendation based on level III evidence). A of haemostatic are available for the treatment of bleeding in patients with haemophilia A and inhibitors. Patients with inhibitors of < 2 BU/ml will to increased doses of human factor It is common clinical that inhibitors of up to 5 BU/ml may be by doses of human factor VIII, although other may be more the inhibitor titre is in of 5 BU/ml, human factor VIII is may not with the patient's antibody to human A retrospective of patients a median inhibitor to porcine of of these patients had no at to this product in the Bethesda assay (Hay et al, has been reported in up to of & 1984; et al, 1984; et al, may the use of porcine less frequently than human after of & 1984; et al, 1984; et al, Specific inhibitors may also following regular replacement therapy in a of patients who have a specific A post-infusion in is common after but is usually and et al, 1984; et al, et al, replacement therapy may be associated with a in This has been to by porcine factor et al, It has been suggested that with this product may an for for the clinical that patients with very high inhibitors may to porcine factor VIII in the absence of a measurable factor VIII increment et al, 1998). of but are usually minor or in degree and the administration of doses et al, 1984; et al, et al, patients have with every infusion and may not be treated regularly with this product (Hay et al, The risk of and the effect on are reduced by the administration of porcine factor VIII by a of administration that should also have pharmacokinetic et al, 1993; Although porcine is not most porcine are not and have not been to to concentrates varying of the IX and It is not how they but it is to be as a result of small of activated and are effective in approximately 50% of et al, VIII inhibitor activity and have a controlled during manufacture and so higher levels of activated was to be more effective than in a controlled with response of and respectively et al, In a controlled was to be no than a et al, with have been reported to be as high as & et al, 1997). et al reported that had controlled bleeding after of surgical of 75% of and of