Feto-maternal alloimmune thrombocytopenia (FMAIT) is due to the destruction of fetal platelets by a maternal platelet specific alloantibody caused by feto-maternal incompatibility, usually for HPA-1a. The most serious complication is intracranial haemorrhage (ICH) which may cause death or neurological sequelae in up to 25% of cases ( Mueller-Eckhardt et al, 1989 ; Kaplan et al, 1996 ). ICH has been observed whatever the platelet alloantigen implicated ( Waters et al, 1991 ).
The recurrence of FMAIT is estimated to be > 80%. In subsequent pregnancies at risk the fetus will be at least as severely affected as the previous sibling. Antenatal diagnosis of fetal thrombocytopenia by percutaneous umbilical blood sampling (PUBS) has revealed that severe thrombocytopenia (< 30 × 109/l) may occur as early as 20 weeks of gestation ( Kaplan et al, 1988 ). The reported frequency of spontaneous ICH in utero is about 10%, although the exact incidence is not known nor is the most likely time of occurrence. In most reported cases ICH has occurred between 30 and 35 weeks, but there are reports of ICH occurring before 20 weeks, and as early as 14–16 weeks ( Giovangrandi et al, 1990 ; Waters et al, 1991 ). There is no spontaneous correction of thrombocytopenia during pregnancy. On the contrary, serial fetal platelet counts show that the platelet count falls as gestation progresses ( Kaplan et al, 1988 ). Antenatal management is therefore aimed at maintaining a satisfactory fetal platelet count and preventing the complications of thrombocytopenia.
The optimal antenatal therapy is uncertain. Available options include maternal therapy with intravenous immunoglobulin (IvIgG), corticosteroids or a combination of both; fetal therapy with direct injection of IvIgG; and in utero platelet transfusions either weekly or immediately before delivery. This report is concerned with maternal therapy. The appropriate use of fetal platelet transfusions was not addressed in this report. Consequently this study excluded the more seriously affected cases who received repeated platelet transfusions as the sole or salvage therapy ( Murphy et al, 1994 ).
There are conflicting reports of the effectiveness of maternal therapy ( Murphy et al, 1994 ; Kroll et al, 1994 ; Bussel et al, 1996 ). Pharmacological studies are particularly difficult in fetuses. The fetal blood is of limited access since cordocentesis cannot easily be repeated. The differences in outcome could result from different treatment protocols, from subtle differences in the therapeutic products, especially for IvIgG, as different brands have been used, or from the evaluation of the results. This evaluation is complex, especially as the published reports are not based on comparative methodology or randomized controlled trials. Such trials are difficult to carry out in a single centre, considering the incidence of this condition (1/1000 live births) ( Durand-Zaleski et al, 1996 ), the risk of fetal damage and ethical issues, including the legal relationship between the mother and her fetus. Collaboration amongst European groups specializing in the management of FMAIT has made it possible to collect more information by setting up a European registry to evaluate antenatal therapy, based on each centre analysing the clinical data in the same way.
This report analyses the outcome of 37 cases of FMAIT linked to anti-HPA-1a immunization managed by maternal therapy from 1989 to 1994. Only cases who followed a similar protocol were included, and each centre decided on the choice of therapy. In 27 cases the mother received weekly infusions of IvIgG (1 g/kg) given slowly, and 10 cases were treated with corticosteroids (0.5 mg/kg/d). The outcome was assessed by changes in the fetal platelet count and the presence or absence of silent or overt ICH during pregnancy and delivery. Whenever platelet transfusions were given, the platelet counts to be considered were those obtained before the transfusion. The final platelet counts were those obtained at the end of the maternal therapy. When serial platelet counts were obtained, a regression analysis was calculated from the platelet counts at different gestation times to assess the trend of the fetal platelet counts.
Patients were allocated to three groups according to outcome as follows: (i) Success of therapy: a significant increase in the fetal platelet counts and no ICH. (ii) Plateau: no significant change in the fetal platelet counts during therapy and no ICH. (iii) Failure of therapy: a significant fall in the fetal platelet counts and ICH in some cases.
The response to therapy is shown in Fig 1 and summarized as follows: IvIgG was successful in 7/27 (26%) cases and steroids in 1/10 (10%) cases. Therapy was unsuccessful in 9/27 (33%) cases treated with IvIgG and in 7/10 (70%) cases treated with steroids. In 11/27 (41%) cases treated with IvIgG and 2/10 (20%) cases treated with steroids there was no significant change in the fetal platelet count (the plateau group).
For each group we compared different clinical features including maternal age, number of siblings, number of affected siblings and among them those with ICH, the platelet count at the first PUBS, the duration of therapy, the platelet count at the end of therapy, and the duration of gestation. The analysis for maternal therapy with IvIgG is shown in 1 Table I. There were too few cases treated with corticosteroids (only 10 cases) to be analysed for each feature in this way.
Nothing in the clinical data of the patients studied was indicative of the expected outcome of therapy, apart from the previous sibling history. In the therapy success group (seven current pregnancies), out of 10 previous siblings, eight were affected and of these one had ICH and one had died. In the plateau group (11 current pregnancies), out of 16 previous siblings, 15 were affected and of these only two had ICH. In the therapy failure group (nine current pregnancies), out of 19 previous siblings, 14 were affected, four with ICH and three had died. Although the numbers are not statistically significant, there were proportionally more severely affected siblings (ICH or death) in the therapy failure group (29% ICH, 21% death) than in the plateau (13% ICH, no deaths) or success (12% ICH, 12% death) groups.
In the therapy failure group, haemorrhagic complications were observed in three of the current pregnancies, with a cord haemorrhage in one causing abortion, and ICH in two leading to neurological sequelae in one of these and death in the other. There was also one fetal death from unknown causes.
The higher fetal platelet counts at the first PUBS for the therapy failure group ( Table I) probably reflected the earlier timing of the first PUBS, because these pregnancies were considered at high risk from previous sibling histories. Five cases were considered for antenatal management, notwithstanding normal fetal platelet counts at the first PUBS, because of the severity of FMAIT in a previous sibling, and are indicated by asterisks in Fig 1. There were four such cases in the failure group (150, 161, 169, 187 × 109/l), and one in the plateau group (185 × 109/l). This study confirmed that the fetal platelet counts of an incompatible fetus could fall during pregnancy, even when the platelet count was normal at the first PUBS. This emphasizes the need to assess the fetal platelet count towards the end of pregnancy in cases with affected siblings, even when the first fetal platelet count of an incompatible fetus appears to be satisfactory, so that delivery can be protected with a fetal platelet transfusion if necessary.
From these data, it is difficult to draw firm conclusions about the effectiveness of either therapy. There was a satisfactory response to maternal treatment with IvIgG in 26% of cases and with steroids in 10% of cases. The significance of the plateau of the fetal platelet counts during pregnancy, which accounted for 41% of cases treated with IvIgG and 20% of cases treated with steroids, is uncertain. It may indicate a stabilization of the thrombocytopenia and hence a beneficial effect of therapy. On the other hand, it may indicate the natural course of the platelet count in a low-risk pregnancy, which was not influenced by treatment. Two of the authors have observed some families in which an incompatible fetus was unaffected after two or three previously affected siblings (unpublished observations). Apart from the sibling history, our data do not suggest any predictive parameters for response to therapy. However, a yet unidentified individual factor(s) may be important and may emerge from subsequent data from the European registry.
Prospective and collaborative studies are necessary to establish the optimal antenatal management of high-risk pregnancies. These studies will enable the pooling of controlled clinical data, based on an agreed protocol for antenatal management and for the evaluation of therapy. We do not recommend maternal therapy for antenatal management without assessment of the fetal status by PUBS. Fetal platelet counts must be performed at least twice during pregnancy to assess the response to therapy. This will give an opportunity to modify the therapy and to transfuse platelets if necessary to manage a safe delivery. Due to the risks involved, cordocentesis must be done only by a skilled team and should be covered by a platelet transfusion to prevent bleeding in a severely thrombocytopenic fetus.