Feto-maternal alloimmune thrombocytopenia occurs in about one in 1000 to 2000 pregnancies and, because there is no routine antenatal screening test, the first affected baby is usually diagnosed after birth. Subsequent pregnancies are affected in 75–90% of cases and there is a high risk (20%) of developing intracranial haemorrhage either in utero or after delivery, which may result in death or neurologic impairment in 7% and 19% of cases, respectively1,2. Antenatal treatment in established feto-maternal alloimmune thrombocytopenia is focussed on preventing intracranial haemorrhage, but the optimal treatment is controversial. It has been suggested that intravascular fetal platelet transfusions are necessary in cases where there is a history of severely affected pregnancies3, but there is little consensus about the efficacy of maternal intravenous gammaglobulin (IVIgG) therapy during pregnancy4. Maternal IVIgG treatment has gained greater acceptance in North American centres, where studies have demonstrated better results5–7 compared with European centres8–11.
This report describes the clinical outcome of six pregnancies over seven years in a woman alloimmunised against HPA-1a. The first pregnancy ended in miscarriage before 15 weeks because of intrauterine death. The third and fourth pregnancies also resulted in miscarriage. The second pregnancy, when HPA-1a antibodies were first identified, was also complicated by autoimmune thrombocytopenia and resulted in a brain-damaged infant. In the fifth and sixth pregnancies (one with and one without fetal platelet infusions), maternal IVIgG was effective in the antenatal treatment of feto-maternal alloimmune thrombocytopenia and healthy infants were delivered by caesarean section. IVIgG administration appeared to reduce the titre of HPA-1a antibodies during the course of these two pregnancies.
Outcome of the second pregnancy
A 25 year old woman had epistaxes and a platelet count of 10 × 109/L due to idiopathic thrombocytopenic purpura. Laboratory investigation revealed the presence of maternal platelet-specific IgM autoantibodies and HPA-1a alloantibodies. The mother was genotyped as HPA-1b/1b. The mother received steroids for her idiopathic thrombocytopenic purpura antenatally and IVIgG at 19 and 27 weeks of gestation. After IVIgG treatment, the maternal platelet count increased on both occasions. The infant was delivered by caesarean section at 34 weeks because hydrocephalus had been detected. An infant phenotyped as HPA-1a (+) was born with brain damage caused by a major intracranial haemorrhage. The infant's platelet count at birth was 15 × 109/L. After delivery, the mother was successfully treated for her idiopathic thrombocytopenic purpura by splenectomy. Her platelet count became normal and the platelet-specific IgM autoantibodies disappeared. However, HPA-1a antibodies were still detectable in the maternal circulation after splenectomy. The mother attended the Bristol Fetal Maternity Unit during her fifth and sixth pregnancies.
Outcome of the fifth pregnancy
Amniocentesis was performed at 19 weeks of gestation, because her partner's genotype was HPA-1a/1b. The fetus was found to have inherited HPA-1a from the father. Cordocentesis was performed at 22 weeks of gestation to determine the fetal platelet count. The first fetal platelet infusion was given at 24 weeks. Maternal obesity made fetal transfusion and blood sampling technically difficult and in an attempt to ameliorate the disease, maternal IVIgG (0.4 g/kg, three times a week) infusions were given between 24 and 29 weeks of gestation. Five platelet transfusions were administered to the fetus at intervals of 7–10 days during this period. A healthy infant with a platelet count of 127 × 109/L was delivered by caesarean section at 30 weeks of gestation. Postnatally, there was no evidence of haemorrhage in the infant (Table 1).
|Pregnancy number and comment||Gestational age or age post-delivery (pd) in weeks + days||IVIgG treatment||Platelet transfusions (days since last transfusion/sampling)||Fetal/infant platelet count (×109/L)||Maternal HPA-1a titre|
|2. Two doses of IVIgG given for maternal ITP, not for FMAITP||Delivery at 34 weeks||15|
|7 + 6 days (pd)||128|
|5. First pregnancy treated by IVIgG||18 + 3||64|
|22 + 2||220||NA|
|24 + 3||Yes||Transfusion given||19||499|
|25 + 3||Yes||Transfusion given (7)||40||515|
|26 + 4||Yes||Transfusion given (8)||43||463|
|28 + 0||Yes||Transfusion given (10)||35||464||8|
|29 + 2||Yes||Transfusion given (9)||68||294|
|Delivery at 30 weeks||127|
|2 + 2 days (pd)||32|
|6. Second pregnancy treated by IVIgG||16 + 6||Yes||16|
|21 + 5||Yes||Transfusion given||70|
|22 + 2||Yes||16|
|23 + 4||Yes||Not given (13)||140||NA|
|27 + 5||Yes||Not given (29)||139||NA|
|28 + 2||Yes||4|
|29 + 2||Yes||4|
|31 + 5||Yes||Not given (28)||131||NA|
|Delivery at 34 weeks||211|
|1 + 3 (pd)||32|
|6 + 5 (pd)||32|
|11 + 4 (pd)||32|
Outcome of the sixth pregnancy
Amniocentesis was performed at 15 weeks of gestation and genotyping revealed that the fetus was HPA-1a/1b. Maternal IVIgG treatment (0.5 g/kg, two times a week) was given between 16 and 31 weeks of gestation. Cordocentesis was first performed at 21 weeks, and because the procedure was again technically difficult, a small platelet volume (5 mL) was transfused while the platelet count, which was found to be 70 × 109/L, was being determined. Cordocentesis was performed at 23, 27 and 31 weeks of gestation, when the platelet count was greater than 130 × 109/L and no further platelet infusions were performed. A healthy infant with a platelet count of 211 × 109/L was delivered at 34 weeks by caesarean section. There were no neonatal complications (Table 1).
Platelet immunology investigations
Platelet-specific antibodies were detected using an immunofluorescence test12 and the monoclonal antibody immobilisation of platelet antigens (MAIPA) assay13, together with a panel of HPA-typed platelets.
The titre of HPA-1a antibodies in samples obtained at the end of the second pregnancy and during the course of both the fifth and sixth pregnancies was determined in a MAIPA assay using platelets from a single HPA-1a/1b platelet donor. The samples were tested retrospectively in the same assay to eliminate inter-assay variation. The HPA-1a antibody titre was 128 at the end of the second pregnancy and decreased from 64 to 8 during the fifth pregnancy and from 32 to 4 during the sixth pregnancy. The antibody titre increased after delivery, when IVIgG was withdrawn. Over the three and half year period of study, there was an overall decrease in HPA-1a titre from 128 to 32 (Table 1).
Genomic DNA isolated from mononuclear blood cells was typed for HPA-1 to -5 using a polymerase chain reaction with sequence-specific primers14.
Feto-maternal alloimmune thrombocytopenia is caused by the formation of maternal antibodies to a platelet antigen inherited by the fetus from the father4. The treatment of pregnancies at risk of feto-maternal alloimmune thrombocytopenia is controversial. It has been recommended that serial cordocentesis be performed, combined with platelet transfusions if the fetal platelet count is <50 × 109/L, as this is the level at which haemorrhage is likely15. However, there is a risk of bleeding at cordocentesis if the fetus is thrombocytopenic16, and the procedure must be repeated every 7–10 days because of the short half-life of infused platelets. An alternative or complementary strategy is to infuse IVIgG into the mother during pregnancy. IVIgG maintains or increases fetal platelet counts in some cases. Better results have been obtained in North America compared with Europe. The reasons for this inconsistency are unclear, but it may be partly explained by differences in the assessment of clinical outcome (i.e. the diagnosis of intracranial haemorrhage, determination of fetal and neonatal platelet counts, the source of the IVIgG17 and the dosage and timing of treatments). A failure of transplacental passage of IgG could also explain the failure of the platelet count to increase in certain cases.
In this case report, three pregnancies (Nos. 1, 3 and 4) were terminated after intrauterine death, but it is not known whether platelet antibodies contributed to the mortality. In the second pregnancy, HPA-1a and platelet autoantibodies were detected in the maternal circulation and IVIgG treatment was given on two occasions to increase the maternal platelet count. This was successful, but the treatment was insufficient to prevent thrombocytopenia and intracranial haemorrhage in the infant. During the fifth pregnancy, IVIgG was given three times weekly from week 24, together with five fetal transfusions at 7–10 day intervals between weeks 24 and 29. Good increases in the platelet count were obtained after each platelet infusion and the interval between transfusions tended towards the maximum expected from the half-life of transfused platelets. This combined treatment resulted in a healthy, non-thrombocytopenic infant. In the sixth pregnancy, despite the fact that alloimmune thrombocytopenia usually becomes more severe in subsequent pregnancies4, no transfusions were necessary because the platelet count remained above 130 × 109/L from 23 weeks onwards. In this pregnancy, the IVIgG treatment was started earlier (16 weeks) and at a higher dose than in the fifth pregnancy. This pregnancy resulted in a healthy, non-thrombocytopenic infant. These findings suggest that IVIgG can be effective either as an adjunct or possibly as an alternative to multiple fetal platelet transfusions.
In the fifth and sixth pregnancies there was a decrease in HPA-1a titre during IVIgG treatment, which was followed by an increase after stopping the treatment and delivery. It is generally accepted that part of the action of IVIgG is to block Fc receptors in the fetal reticulo-endothelial system. However, an additional mechanism may be to reduce the titre of maternal platelet alloantibodies, possibly by a negative feedback inhibition of plasma cell antibody production. The reduction in HPA-1a titre is greater than that anticipated due to the haemodilution of pregnancy18.
Platelet-specific antibodies are able to cross the placenta from 14 weeks of gestation19, and while most intracranial haemorrhages occur in the third trimester, they have been reported as early as 20 weeks20. Therefore, in high risk cases, it is reasonable to start IVIgG before this, with the first fetal blood sample taken between 20 and 24 weeks21. It is important that cordocentesis is repeated at intervals to determine the fetal response, with HPA-compatible platelets being available if the fetus is thrombocytopenic. Further research in the UK involving other fetal medicine centres is needed to determine the optimal management of feto-maternal alloimmune thrombocytopenia, which is likely to include a combination of maternal IVIgG treatment and fetal platelet transfusions.
The authors would like to thank Sue Rogers, Becky Griffiths and Sharon Scourfield (IBGRL) for technical assistance in performing the platelet immunology investigations and Dr Mike Murphy (NBS-Oxford) and Professor Adrian Newland (Royal London Hospital) for clinical advice. We are indebted to Dr Paul Metcalfe (National Institute of Biological Standards and Control, Herts.) for the provision of primers for the PCR-SSP technique.