Prediction of IVIG treatment efficiency in fetal/neonatal alloimmune thrombocytopenia

To the editor:

Fetal/neonatal alloimmune thrombocytopenia resulting from specific maternal immunization against human platelet antigens (HPAs) is the most frequent cause of severe isolated thrombocytopenia (platelet counts <50 × 10⁹/L) in maternity wards, and its most destructive consequence is intracranial hemorrhage (20% to 25% of reported cases of fetal/neonatal alloimmune thrombocytopenia).¹  In the absence of routine antenatal screening, diagnosis is mostly carried out when the fetus or newborn is already affected. Because of high recurrence and greater severity of this syndrome in subsequent pregnancies with incompatible fetus, there is a consensus on antenatal management with maternal first-line therapy using intravenous immunoglobulins (IVIG) ± steroids.^2-4  Fetal platelet count is mostly unknown during pregnancy because of the risk of fetal blood sampling. Consequently, maternal parameters have been investigated to evaluate fetal status and its response to maternal therapy. Since 2004, our laboratory has carried out a prospective study on maternal anti-HPA-1a alloantibody concentration follow-up during managed pregnancies. Since our previous publication,⁵  we have increased our cohort (55 HPA-1bb women). Pregnancies were managed with IVIG from 20.5 (median) ± 2.8 weeks of gestation (1 g/kg per week dose), and steroids during the last trimester for 38 of these 55 women (prednisone 0.5 mg/kg per day). Sequential maternal anti-HPA-1a quantifications have been carried out using the gold-standard method of monoclonal antibody-specific immobilization of platelet antigens⁶  on a total of 329 samples (median of 5 ± 2 sera per pregnancy) from the beginning of IVIG therapy to delivery (1 sample every 3 to 5 weeks). There were 4 twin pregnancies and 59 babies were born at 37 (median) ± 1.6 weeks of gestation. For each pregnancy, antibody concentrations were analyzed by calculation of the area under the curve (AUC) weighted by the weeks between the first and the last quantification.^5,7  The most striking result was the correlation between AUC and newborn platelet count at delivery: low AUC was correlated with a safe newborn platelet count (>50 × 10⁹/L) and high AUC with severely affected newborns needing postnatal treatment (platelet transfusion with/or without IVIG). The cohort presented here allowed the establishment of an AUC threshold of 23 IU/mL (P < .0001; sensitivity, 76.19%; specificity, 78.95%) (Table 1) with a good negative predictive value (85.71%): AUC can help clinicians predict a fetal response to maternal therapy. If we consider the obstetric history of the mother (IVIG-treated pregnancy at gestation 2 or >2), predictive values of AUC for response to therapy were better at second gestation than for multigestational women (Table 1). This larger series confirms that multigravida is an obvious risk factor in therapy failure.⁵  These new data confirm our preliminary results concerning the predictive value of maternal anti-HPA-1a alloantibody follow-up for fetal therapy response, more specifically for second-gestation women. This noninvasive parameter will help clinicians manage delivery in high-risk pregnancies. According to the AUC measured prepartum with the sequential samples collected during pregnancy, we may consider (1) vaginal delivery with low risk of severe thrombocytopenia when AUC is below 23 IU/mL or (2) cesarean section when there is a higher risk of severe thrombocytopenia when AUC is above this threshold.