Acquired factor VIII inhibitors in pregnancy: data from the Italian Haemophilia Register relevant to clinical practice
* Dr F. Baudo, Thrombosis and Hemostasis Unit, Niguarda Hospital, P.za Ospedale Maggiore 3, 20162 Milan, Italy.
Objective To review the clinical problems related to the inhibitor of factor VIII (FVIII) in pregnancy.
Design Retrospective analysis.
Setting Haemophilia and haemotology centres.
Population Patients registered and followed up at the centres.
Methods Data were collected from the Italian Haemophilia Register of acquired FVIII inhibitor.
Results Twenty of 96 cases with FVIII inhibitor were identified postpartum. The time of appearance was 3–150 days postpartum. All but one of the cases were idiopathic; 11/20 patients required blood transfusions. In six patients, the inhibitor was identified because of surgical bleeding, four after hysterectomies carried out because of postpartum haemorrhage. A prolonged activated partial thromboplastin time was present in all women in whom the test was carried out. Nine women did not require treatment because the bleeding was mild; in 11 patients bleeding was promptly controlled by different therapeutic modalities. Immunosuppressive therapy was used to suppress the inhibitor. The majority of the patients who achieved complete remission received steroids; in 6/6 patients who relapsed, a second remission was obtained with combined therapy.
Conclusions In a review of 20 pregnancies with FVIII inhibitor, over a 15 year period, bleeding was controlled in all cases with no fatalities. Correct evaluation of coagulation screening tests, in particular, activated partial thromboplastin time, is essential.
The occurrence of FVIII (anti-haemophilic factor) antibodies in the general population is a rare event: several surveys reported an incidence of 0.2–1 × 106/year1,2. The incidence increases with age 50–60% of the cases occurring over the age of 503,4. Overall, the incidence is higher in men (52.7%), but in the younger age group, the incidence is higher in women because of the relation to pregnancy3. Acquired haemophilia occurs in association with autoimmune conditions (systemic lupus erythematosus, rheumatoid arthritis, asthma) and neoplastic diseases, drug hypersensitivity and pregnancy. However, 50% of cases are idiopathic3–5. Common sites of bleeding are skin (large ecchymoses), mucosa (epistaxis, gengivorrhagia, metrorrhagia), muscles and retroperitoneum. Bleeding can arise suddenly and spontaneously or after procedures (intravenous catheter positioning, surgery, intramuscular injections). If bleeding occurs in critical sites, compression problems can ensue. The severity of bleeding is not proportional to the inhibitor titre6. The reported mortality related to the inhibitor is about 10–22%2–5. The prolongation of activated partial thromboplastin time with normal prothrombin time is the hallmark of laboratory diagnosis. The objectives of therapy are control of bleeding and suppression of the autoantibodies.
The inhibitor may occur in otherwise healthy pregnant women without a previous history of bleeding. It may disappear spontaneously or persist for months or years with a consequent high risk of bleeding3,7. In 1999, the Italian Haemophilia Centres carried out a survey of cases of acquired FVIII inhibitor registered in the last 15 years. The data relevant to this clinical problem in pregnancy are the object of this report.
A questionnaire was mailed to 42 Italian Haemophilia Centres to collect data referring to patients reported during the previous 15 years. The following information were requested:
- •the primary condition (if present); the initiating cause of bleeding (e.g. trauma, surgery, other events) or its spontaneous occurrence; the site of bleeding and its severity evaluated by the haemoglobin level and by the transfusion requirement;
- •the type (anti-FVIII or anti-FIX), the titre of the inhibitor and the anamnestic response if present;
- •the therapy administered to control the bleeding; agents used, modality of administration and results;
- •the immunosuppressive therapy; drugs and modality of their administration, side effects and results. Complete remission is defined as normal level of FVIII and no detectable inhibitor; partial remission is defined as an inhibitor titre <10 Bethesda unit (BU) or a decrease of 50% if the baseline inhibitor titre was <10 BU; failure is defined as no change or an inhibitor titre >10 BU or a decrease less than 50% of the baseline values.
Twenty centres contributed information on a total of 96 patients; 15% of the forms were incomplete. Therefore, the data do not refer to all of the patients. Forty-five cases (47%) were idiopathic; 20 cases (21%) were related to pregnancy. The characteristics of the pregnant women and the bleeding problems are detailed in Table 1. The median follow up was eight years (0.5–15). Sixteen women were primiparas. The inhibitor was identified on the occasion of bleeding in all women. The survey did not yield information on the newborns. No recurrence of the inhibitor was reported in the four women who had a successive pregnancy. The median time to the onset of significant bleeding and the identification of the inhibitor after delivery was 60 days (range 1–150 days); in a few women, a previous mild bleeding was overlooked and was not investigated. Nineteen cases were idiopathic and one patient had a concomitant autoimmune disease (thyroiditis). The activated partial thromboplastin time was prolonged in all women in whom the test was reported; the median titre of inhibitor was 8.5 BU/mL (range 1.6–140).
Table 1. Characteristics of the women and site and intensity of bleeding.
|Age (years)a||30 (20–40)|
|Relation to pregnancy (1st/≥2nd)b||16/4|
|Time to bleeding from delivery (days)a||60 (1–150)|
|Idiopathic/autoimmune disease (patient no.)||19/1|
|Prolonged partial thromboplastin time (patient no.)||15c|
|Inhibitor titre (BU/mL)a||8.5 (1.6–140)|
|Site of bleeding (patient no.)|
|Postpartum haemorrhagia (subsequent hysterectomy)||3|
|Caesarean section (subsequent hysterectomy)||1|
|Postpartum haemorrhagia and/or haematomata||14|
|Intensity of bleeding (patient no.)|
|Haemoglobin at presentation (g/dL)a||8.7 (5–13)|
In 6 out of 20 women (30%), the inhibitor was identified because of bleeding during or after an intervention. In two patients the bleeding occurred after a tooth extraction and an intramuscular injection, respectively; in four cases hysterectomy was carried out because of uncontrolled haemorrhage (three cases after vaginal, one case after caesarean delivery). At diagnosis the median haemoglobin level was 8.7 g/dL (range 5–13); 11 women required blood transfusions.
In 11 women bleeding was controlled with different therapeutic modalities: human or porcine FVIII; activated recombinant FVII (rFVIIa); high dose immunoglobulins combined with activated prothrombin complex concentrate or porcine FVIII or activated recombinant FVII; plasmapheresis and porcine FVIII. Nine women, who did not require treatment, experienced a mild and delayed bleeding. The characteristics of the patients in relation to therapy are outlined in Table 2. In two women who did not receive the anti-haemorrhagic therapy, the inhibitor disappeared without immunosuppression after two and four months, respectively. Eighteen women, including the seven patients who did not require treatment for bleeding, were treated with immunosuppressive therapy. The initial treatment was prednisone alone in 10, and prednisone combined with other agents (cyclophosphamide, azathioprine or high dose immunoglobulins) in eight women.
Table 2. Characteristics of the women in relation to the treatment for bleeding control and response to immunosuppressive therapy.
|Therapy to control bleeding||Yes||No|
|Inhibitor titre (BU/mL)a||4 (1.7–140)||11 (2.5–33)|
|Time to bleeding (relation to partum) (days)a||8.5 (1–150)||90 (1–150)|
|Type of bleeding||Haemorrhagia, haematomata||Mild ecchymoses|
|Haemoglobin (g/dL)a||7.8 (5–10.7)||11.2 (7–13)|
|No. of patients requiring transfusions||9||2|
|First complete remission||10||4|
|First partial remission||1||2|
|Second complete remission||3c||3c|
Complete remission and partial remission were obtained in 14 (78%) and 3 (16%) women, respectively (response rate 94%). The median time to the first complete remission was 2.7 months (range 0.5–36) and the median time of the treatment was 4.0 (range 1–18) months. In two women with partial remission, the inhibitor persisted at a low titre (1.4 and 3 BU, respectively); treatment was discontinued and the inhibitor disappeared after 10 and 36 months, respectively. In one woman the inhibitor was still present with the titre unchanged (2.5 BU) without bleeding after a follow up of 48+ months. Immunosuppressive therapy was carried out for six months. Six women relapsed after a median time of 16 months (range 6–24 months) and were rescued with combined additional treatment (steroid + cyclophosphamide or azathioprine). The outcome of immunosuppressive therapy is similar in the women who did or did not receive anti-haemorrhagic therapy. The time to the first complete remission was independent of the baseline inhibitor titre: median 3 months (range 0.5–18) in eight patients with inhibitor titre <10 BU/mL and median 3.5 months (range 1–36) in six patients with inhibitor titre >10 BU. The inhibitor titre and the time to response were not different in the women who relapsed (median inhibitor titre 6.5 BU, range 3.2–140 and median time to response 2.7 months, range 2–36) and in the women who did not relapse (median inhibitor titre 8.0, range 1.6–32 and median time to response 2.0 months, range 0.5–15). No fatalities were reported.
Pregnancy is a frequent concomitant condition in the FVIII inhibitor syndrome (in our survey 21% of cases). Other studies reported a lower incidence (7–11%)2–4,6. The total number of cases in our survey is lower than expected, probably due to under-referral to the centres; the results of our survey could therefore be biased. The majority of the postpartum cases in our series is idiopathic (19/20). The inhibitor occurred in the first pregnancy in 16 out of 20 women and did not recur in four women who had a successive pregnancy. The same observations were reported by others7–10 but in the survey of Solymoss11 the inhibitor recurred in the three women who had successive pregnancies. The inhibitor was identified because of the occurrence of bleeding in 10 women in the peripartum period and in 10 women after 30 days from delivery. In the series of Solymoss11, the bleeding occurred prepartum in two women, within three days from delivery in three and within 3–12 months in nine. In the review of Michiels et al.10, bleeding occurred during pregnancy in 1, after abortion in 1, immediately after delivery in 4, within 4 months in 8, after a period of 4 months to over 12 months in 27 women, respectively. The time of the occurrence of bleeding is quite variable. The inhibitor is in general identified on occasion of overt bleeding; the time of the development of the inhibitor in the absence of bleeding signs cannot be retrospectively determined.
Nine patients (45%) did not require blood transfusions; the bleeding was mild with a stable haemoglobin and was self-limiting. The onset of bleeding in six of these women was delayed (range 60–150 days) at variance with the patients who required therapy. In the survey of Solymoss11, 3 out of 14 patients did not require therapy to control the bleeding. In our survey, the inhibitor titre did not correlate with the intensity of bleeding but correlation was reported by other investigators7,11.
In two women, the bleeding occurred after an intervention (tooth extractions, intramuscular injection). Four women underwent hysterectomy because of postpartum haemorrhage. In 14 women, including the four who underwent hysterectomy, the prolonged activated partial thromboplastin time was already known, but the inhibitor was recognised after the onset of bleeding. It must be emphasised that the prolonged activated partial thromboplastin time was overlooked. With an adequate diagnosis and replacement treatment, the hysterectomies could have been avoided. Mortality is low. No fatalities were reported in our and in the Solymoss series; 10% (4/40) in the Michiels et al.10 survey.
Inhibitors may cross the placenta and may persist up to three months in the neonate usually without bleeding complications12–14. However, Ries et al.15 reported a case of intracranial haemorrhage in a neonate. These data suggest that delivery should be managed as in haemophilia16.
The immunosuppressive treatment with steroids alone or in combination with other agents was the preferred treatment (17/18). The response rate was high (94%), the relapse rate was also high (42%) but all the women were rescued. The immunosuppressive therapy may shorten the time to response without influencing the response rate. Hauser et al.7 reviewed the data in the literature comparing the immunosuppressive therapy versus no treatment. Time to response was shorter in the treated women but the overall response rate was not different. Similar results were reported by Michiels et al.10. These studies must be considered with caution for a number of reasons. They are retrospective, refer to a small number of patients with heterogeneous characteristics and have no predefined criteria for treatment. Nevertheless, steroids may be considered the treatment of choice16. Only two women did not receive therapy either for bleeding or immunosuppression. Eighteen women received immunosuppressive therapy. As a general consideration, the final outcome of the postpartum inhibitor syndrome is favourable, independent of the type of treatment. Therefore, the recognition of the syndrome is of utmost importance. Acquired haemophilia usually occurs in general hospitals. A prolonged activated partial thromboplastin time should not be overlooked. In the presence of an unexplained often severe haemorrhage with abnormal coagulation screening, it is important to seek immediate specialist advice. Because of the rarity of the disorder, the treatment modality and the potential risk of severe bleeding, these patients should be managed in a haemophilia centre or under their supervision17,18.
Francesco Baudo, Francesco de Cataldo.
Contributing haemophilia centres
Bari, Schiavon M; Catanzaro, Santoro R; Firenze, Linari S; Milano Niguarda, Mostarda G; Milano Policlinico, Santagostino E; Pavia, Gamba G; Pescara, Dragani A, Roma, Mazzucconi G; Torino, Schinco P; Vicenza, Castaman G.