Berkowitz, Richard L. MD1; Lesser, Martin L. PhD2; McFarland, Janice G. MD3; Wissert, Megan RN4; Primiani, Andrea4; Hung, Crystal4; Bussel, James B. MD4
Alloimmune thrombocytopenia is the most common cause of severe thrombocytopenia in fetuses and term neonates1,2 and the most frequent cause of intracranial hemorrhage in neonates born at term.2,3 Alloimmune thrombocytopenia is caused by parental incompatibility of a platelet-specific antigen, with maternal sensitization to the antigen in question resulting in fetal thrombocytopenia, which is often severe.4,5 Because population screening for this disorder is not routinely performed, couples are almost always identified to be at risk for having an affected infant after they have already had one.
Ten percent to 22% of severely affected infants can develop an intracranial hemorrhage,2 and 75% of those bleeds occur before delivery.2,6 Therefore, this disorder should be treated during the antenatal period.7–9 Maternally administered intravenous immunoglobulin (IVIG) has been the most successful therapy used to date in elevating fetal platelet counts to adequate levels and, thereby, preventing intracranial hemorrhage.10 However, previously untreated fetuses with platelet counts less than 20,000/mL3 are unlikely to adequately respond to IVIG alone given at the standard dose of 1 g/kg/wk and are, therefore, presumably at greater risk of developing an intracranial hemorrhage with this treatment than those with higher counts.10 Therefore, we and other authors have advocated determining the fetal platelet count before the initiation of therapy10,11 and then monitoring the fetal platelet response to determine which patients were candidates for more intensive therapy with higher doses of IVIG or the addition of daily prednisone to the regimen or both.10–12 Unfortunately, the only way to currently obtain the fetal platelet count is to perform fetal blood sampling, which is an invasive procedure that is associated with considerable perinatal morbidity and mortality.10,11,13,14
This report describes a randomized multicenter study to evaluate the effectiveness and safety of two antenatal treatment regimens designed to optimally protect fetuses against having an intracranial hemorrhage resulting from alloimmune thrombocytopenia while minimizing the risks associated with fetal blood sampling. The study was limited to “standard-risk” patients, who were defined as women with documented alloimmune thrombocytopenia who had not delivered an infant with an intracranial hemorrhage in a prior pregnancy. Patients who previously had an affected fetus or neonate with an intracranial hemorrhage were considered to be “high-risk” because they are known to require more intensive therapy and were treated on a separate protocol (Hung C, Wissert M, Primiani A, Berkowitz R, McFarland JG, Bussel JB. Management of fetal alloimmune thrombocytopenia: treatment of 30 cases in which the previous sibling suffered an intracranial hemorrhage [abstract]. ASH Annual Meeting Abstracts. Blood 2006;108:1080).
PARTICIPANTS AND METHODS
Between May 25, 2001, and June 21, 2006, 73 women with documented alloimmune thrombocytopenia whose prior affected children had not suffered an intracranial hemorrhage were enrolled in this randomized prospective study in 35 institutions throughout the United States and at one Canadian center. As in prior studies performed by our group, it was necessary to recruit patients from all over the country because of the relative rarity of the disorder. A common protocol was accepted by the institutional review board of each participating institution, and monitoring was performed by periodic phone calls from the coordinating center with the help of the home nursing services performing the IVIG infusions. Randomization of treatment arms was performed by using computer-generated random numbers with permuted blocks to achieve balance in the design. The process was conducted by having the investigator’s office call the biostatistics office to receive the treatment assignment, with immediate fax confirmation. Cordocentesis procedures were either performed by a local perinatologist or referred to New York City for performance by one of the authors (R.L.B.).
Women were diagnosed as having alloimmune thrombocytopenia if they were shown to have an incompatibility to a Human Platelet Antigen (HPA) on the father’s platelets and to have circulating antibodies to that antigen. If the father was a heterozygote for the antigen in question, the HPA genotype of the fetal platelets was determined by testing cells obtained via amniocentesis during the current pregnancy.
Patients with documented alloimmune thrombocytopenia carrying a fetus whose platelets contained the offending antigen were initially stratified according to whether they had delivered an infant that had suffered an intracranial hemorrhage during a prior pregnancy. Those that had not were randomized into one of two therapeutic groups, and they constitute the study population described in this report. Women whose prior infant had developed an intracranial hemorrhage in utero or during the neonatal period were treated differently and are not included in this study.
Patients in the study group were randomized to receive either IVIG 2 g/kg/wk (group A) or IVIG 1 g/ kg/wk plus prednisone 0.5 mg/kg/d (group B) starting at approximately 20 weeks of gestation. A single cordocentesis procedure was performed at approximately 32 weeks after the prophylactic administration of betamethasone. The technical aspects of performing this procedure have previously been described.10 “Salvage therapy” was initiated if the fetal platelet count was found to be less than 30,000/mL3, or the procedure could not be performed. Salvage therapy for patients in group A consisted of adding prednisone 0.5 mg/kg/d to their regimen, while that for group B was an increase in IVIG to 2 g/kg/wk in addition to the prednisone they were already taking. A salvage failure was defined as a birth platelet count less than 30,000/mL3 (Fig. 1).
The primary outcome variable was the development of a fetal or neonatal intracranial hemorrhage. Secondary outcome variables were fetal and birth platelet counts, gestational age at delivery, problems associated with fetal blood sampling, and complications of medical therapy as reported by patients on a questionnaire administered after the delivery.
Data are summarized by using frequencies, percentages, means, and ranges. Outcome rates were estimated by using exact binomial confidence intervals (CIs) and compared by using the Fisher exact test.
Thirty-seven pregnant women, including three carrying twins, were randomized to group A (IVIG 2 g/ kg/wk) and 36, including one woman with twins, to group B (IVIG 1 g/kg/wk plus prednisone 0.5 mg/kg/d). The demographic, delivery, and platelet data for both groups are listed in Tables 1 and 2.
One neonate in each group suffered an intracranial hemorrhage during the neonatal period (P=1.00, 95% CIs 0.1–14% and 0.1–15%), and neither was due to treatment failure. The infant in group A was delivered vaginally at 37 weeks with a birth platelet count of 133,000/mL3 and subsequently developed a grade 1 subependymal hemorrhage. The infant in group B developed a grade 1 subependymal hemorrhage, which was detected at 1 week of life after being delivered by cesarean at 32 weeks with a birth platelet count of 197,000/mL3.
The average platelet counts at the time of fetal blood sampling were 121,600/mL3 and 116,100/mL3, and the average birth platelet counts were 169,400/mL3 and 134,000/mL3 for groups A and B, respectively. Five patients (14%) in group A delivered infants with birth platelet counts less than 50,000/mL3, compared with 4 (11%) in group B (P=1.00, Fisher exact test, 95% CIs 5–29% and 3–26%). Ten of 37 (27%) patients in group A and 6 of 36 (17%) in group B required salvage therapy (P=.4, 95% CIs 14–44% and 6–33%), and only one of those in each group delivered a child with a birth platelet count less than 30,000/mL3.
Nine fetuses in group A and five in group B had platelet counts less than 30,000/mL3 when sampled at approximately 32 weeks. Two of these women were delivered electively at the parent’s request, and all of the others were placed on salvage therapy. There was one woman in group A and three in group B who had fetal platelet counts between 30,000/mL3 and 50,000/mL3, and none were placed on salvage therapy. The three neonates in group B all had birth platelet counts between 40,000/mL3 and 84,000/mL3, but the infant from group A who had a platelet count of 48,000/mL3 at 32 weeks was born with a platelet count of 14,000/mL3 at 36 weeks.
There were two complications from 39 cordocentesis procedures performed in group A. In both cases sustained fetal bradycardia led to emergency cesarean delivery of the infant. The deliveries were at 33 and 32 weeks, and the birth platelet counts were 104,000/mL3 and 52,000/mL3, respectively. The only complications after 40 fetal blood sampling procedures performed in group B were two cases of premature rupture of the membranes within 24 hours. These infants were both delivered by cesarean at 33 and 37 weeks, with birth platelet counts of 53,000/mL3 and 40,000/mL3, respectively.
Information relating to complications associated with the therapeutic regimens was obtained from 31 patients in group A and 32 in group B. These data are listed in Table 3. It should be noted, however, that nine patients in the former group and four in the latter were placed on salvage therapy, which means that these women received both the higher dose of IVIG and daily prednisone. Table 4 lists the complications reported by those patients in groups A and B who remained on the original therapy to which they had been randomized, as well as those experienced by the 13 women who received salvage therapy.
Only one woman was unable to complete therapy in the arm to which she was randomized. This patient, who was in group B, had a fetal platelet count of 159,000/mL3 at 30 weeks while taking IVIG 1 g/ kg/wk plus prednisone 0.5 mg/kg/d. At that time, however, she discontinued the IVIG because of a severe rash and subsequently delivered an infant with a birth platelet count of 17,000/mL3 at 34 weeks after having been treated with prednisone 1 mg/kg/d alone for 4 weeks.
Alloimmune thrombocytopenia is a disease that has a wide spectrum of severity. Women with this disorder who have delivered an infant in a prior pregnancy that either suffered an intracranial hemorrhage in utero or during the neonatal period are at substantially increased risk of having this occur again, and they require aggressive antenatal therapy that should be started early in gestation.2,6–8,11 On the other hand, women whose prior affected infants did not have an intracranial hemorrhage generally have milder disease, are at lower risk of having a subsequent fetus suffer a bleed in utero, and should it occur, the risk of bleeding before 20 weeks is extremely unlikely.
Our recently published series of 79 patients with alloimmune thrombocytopenia prospectively segregated patients into two categories based on their history and initial fetal platelet counts.10 The high-risk group was defined as having either an affected sibling with an intracranial hemorrhage that occurred during the peripartum period or an initial fetal platelet count before initiating therapy during the current pregnancy of less than 20,000/mL3, while the standard-risk group consisted of patients who did not have a sibling with an intracranial hemorrhage and in whom the initial fetal platelet count in the current pregnancy was greater than 20,000/mL3. Patients in the high-risk group were randomized to receive either IVIG 1 g/ kg/wk alone or in combination with prednisone 1 mg/kg/d. The response to these regimens was found to be substantially different. Eighty-nine percent of those receiving the combination therapy had a satisfactory initial response to therapy, compared with only 35% in the group receiving IVIG alone. Furthermore, in the subset of 22 patients whose initial fetal platelet count was less than 10,000/mL3, a satisfactory response to treatment with IVIG and prednisone was achieved in 82% of the cases, compared with only 18% in those treated with IVIG alone. On the other hand, patients in the standard-risk group were randomized to receive either IVIG 1 g/kg/wk or prednisone 0.5 mg/kg/d, and both of those regimens were found to be efficacious and comparable. Analysis of the 72 patients in that series, who did not have a prior pregnancy complicated by a peripartum intracranial hemorrhage, revealed that the initial fetal platelet count was less than 20,000/mL3 in 46% of the cases and less than 10,000/mL3 in 24%.
These published data clearly indicate that patients with milder disease can be successfully treated with less intensive therapy than their more severely affected counterparts. Therapy with IVIG is expensive, and both IVIG and prednisone can cause adverse maternal adverse effects. Therefore, there are obvious benefits to avoiding overtreatment whenever possible. On the other hand, simply treating all patients with IVIG 1 g/kg/wk will significantly undertreat some patients whose fetuses would then continue to be at risk of having an intracranial hemorrhage in utero.
Stratification and subsequent management of patients in the above-mentioned series required the performance of fetal blood sampling, and the cost of obtaining that information was high. Serious complications occurred in 6% of the fetal blood sampling procedures, and this resulted in the emergent delivery or death in utero of 14% of the 79 infants in that series.10 Other authors have identified similar rates of serious complications associated with fetal blood sampling in patients with alloimmune thrombocytopenia.11,13,14 The fetal-neonatal morbidity and mortality associated with this invasive procedure are clearly substantial, which suggests that providing empiric therapy without performing fetal blood sampling may be safer for infants whose siblings from a prior affected pregnancy had not suffered an intracranial hemorrhage. If that were done, however, it must be re-emphasized that therapy that maximizes platelet response in the most severely affected fetuses will overtreat those with milder disease and be associated with maternal adverse effects that could potentially be avoided if less intensive therapy was administered.
The current study was designed to prospectively identify a group of patients who could be empirically treated with a medical regimen that adequately covers the severely affected fetuses without subjecting the mothers of more mildly affected fetuses to far more therapy than is necessary. Fetal blood sampling was used to evaluate the effectiveness of therapy, but this was not performed until 32 weeks of gestation because, if complications from the procedure were encountered, it was felt that emergent delivery at that time provided the potential for a very good neonatal outcome. This, of course, is less likely to be the case if the procedures were performed much earlier. Intravenous immunoglobulin was given at a dose of 2 g/ kg/wk or at the standard dose of 1 g/kg/wk in combination with daily prednisone to more adequately treat fetuses with initial platelet counts less than 20,000/mL3 who are known to be relatively unresponsive to IVIG 1 g/kg/wk alone. We chose to administer prednisone at a dose of 0.5 mg/kg/d as opposed to twice that amount, which had been our “salvage regimen” in prior studies, because anecdotal experience in previous studies had suggested that taking 1 mg/kg/d of prednisone for several months had resulted in considerable maternal discomfort.
The results of this study are encouraging. The outcomes of both groups were excellent and quite comparable. Statistical analyses of the incidence of intracranial hemorrhage, as well as fetal and birth platelet counts in the two treatment arms, showed that neither is demonstrably superior to the other. A trial to definitively document the equivalence of the treatment arms relative to these parameters would involve hundreds, if not thousands, of patients. In any case, there were no cases of intracranial hemorrhage secondary to thrombocytopenia or life-threatening neonatal consequences of fetal blood sampling because of the gestational age at which it was performed.
The adverse effects associated with the treatment regimens were largely predictable. There was a higher incidence of gestational diabetes and a tendency to more fluid retention, mood swings, insomnia, and jitteriness in patients on prednisone, and of moderate-to-severe fatigue in those on high-dose IVIG alone. Only one patient was unable to tolerate the treatment arm to which she was randomized. This woman developed a severe rash while taking IVIG 1 g/kg/wk plus prednisone 0.5 mg/kg/d and was treated off-protocol with prednisone 1 mg/kg/d alone for the last 4 weeks of her pregnancy. This is the only patient in any of our studies, which include more than 200 pregnant women, who was unable to tolerate IVIG.7,9,10,12,15
Based on the findings of this study, we suggest the following recommendations for managing women carrying a fetus at risk for thrombocytopenia from documented alloimmune thrombocytopenia who have had no prior infants with intracranial hemorrhage:
1. Start empiric therapy with IVIG 2 g/kg/wk alone or IVIG 1 g/kg/wk plus prednisone 0.5 mg/kg/wk as close to 20 weeks as possible, after explaining the adverse effects of both regimens to the patient and her partner. The treatment arms are sufficiently comparable that the choice of regimen can be tailored to the individual patient, although there is a considerable increase in the cost and time of administration when IVIG alone is given at the higher dose. However, if, for example, a woman developed gestational diabetes in a prior pregnancy, IVIG 2 g/kg/wk might be a better option than giving her prednisone throughout the second and third trimesters.
2. Offer all patients fetal blood sampling at 32 weeks and institute salvage therapy for fetal platelet counts less than 50,000/mL3. If fetal blood sampling is not or cannot be performed at that time, empirically institute salvage therapy at 32 weeks, ie, increase the regimen to IVIG 2 g/kg/wk plus prednisone 0.5 mg/kg/d. In this and previous studies, we have used a fetal platelet count of less than 30,000/mL3 as our cutoff for initiating salvage therapy. The recommendation to use 50,000/mL3 as the new cutoff is based on our finding that one fetus in group A of this series with a platelet count of 48,000/mL3 at 32 weeks was not offered salvage therapy, and the mother delivered an infant with a birth platelet count of 14,000/mL3 several weeks later.
3. Allow vaginal delivery only for those patients whose fetuses were demonstrated to have a platelet count greater than 100,000/mL3 at 32 weeks and remain compliant with therapy. This recommendation is based on our uncertainty that any lower value can be relied upon to guarantee that the fetal platelet count will continue to be greater than 50,000/mL3 for 3 or more weeks and our belief that vaginal delivery carries an increased risk of intracranial hemorrhage in severely thrombocytopenic fetuses. An alternative approach is to empirically institute salvage therapy in all patients at 32 weeks and then only perform cordocentesis at 36 weeks for women who want to deliver vaginally. Under those circumstances the morbidity associated with fetal blood sampling should be very low, and vaginal delivery could be offered to patients with fetal platelet counts greater than 50,000/mL3.
Women with a history of intracranial hemorrhage in a fetus or neonate during a prior pregnancy have more severe disease than those included in this study and are at much higher risk of having an affected fetus suffer an intracranial hemorrhage if inadequately treated in utero. Recommendations for the management of those patients will be the subject of a separate report.
1. Hohlfeld P, Forestier F, Kaplan C, Tissot JD, Daffos F. Fetal thrombocytopenia: a retrospective study of 5,194 fetal blood samplings. Blood 1994;84:1851–6.
2. Bussel JB, Zacharoulis S, Kramer K, McFarland JG, Pauliny J, Kaplan C. Clinical and diagnostic comparison of neonatal alloimmune thrombocytopenia to non-immune cases of thrombocytopenia. Pediatr Blood Cancer 2005;45:176–83.
3. Mao C, Guo J, Chituwo BM. Intraventricular haemorrhage and its prognosis, prevention and treatment in term infants. J Trop Pediatr 1999;45:237–40.
4. van Loghem JJ, Dorfmeijer H, van Hart M, Schreuder F. Serological and genetical studies on a platelet antigen (Zw). Vox Sang 1959;4:161–9.
5. Pearson HA, Shulman NR, Marder VJ, Cone TE Jr. Isoimmune neonatal thrombocytopenic purpura: clinical and therapeutic considerations. Blood 1964;23:154–77.
6. Herman JH, Jumbelic MI, Ancona RJ, Kickler TS. In utero cerebral hemorrhage in alloimmune thrombocytopenia. Am J Pediatr Hematol Oncol 1986;8:312–7.
7. Bussel JB, Berkowitz RL, McFarland JG, Lynch L, Chitbara N. Antenatal treatment of neonatal alloimmune thrombocytopenia. N Engl J Med 1988;319:1374–8.
8. Murphy MF, Metcalfe P, Waters AH, Ord J,Hambley H, Nicolaides K. Antenatal management of severe feto-maternal alloimmune thrombocytopenia: HLA incompatibility may affect responses to fetal platelet transfusions. Blood 1993;81:2174–9.
9. Lynch L, Bussel JB, McFarland JG, Chitkara U, Berkowitz RL. Antenatal treatment of alloimmune thrombocytopenia. Obstet Gynecol 1992;80:67–71.
10. Berkowitz RL, Kolb EA, McFarland JG, Wissert M, Primani A, Lesser M, et al. Parallel randomized trials of risk-based therapy for fetal alloimmune thrombocytopenia. Obstet Gynecol 2006;107:91–6.
11. Birchall JE, Murphy MF, Kaplan C, Kroll H; European Fetomaternal Alloimmune Thrombocytopenia Study Group. European collaborative study of the antenatal management of feto-maternal alloimmune thrombocytopenia. Br J Haematol 2003;122:275–88.
12. Bussel JB, Berkowitz RL, Lynch L, Lesser ML, Paidas MJ, Huang CL, et al. Antenatal management of alloimmune thrombocytopenia with intravenous gamma-globulin: a randomized trial of the addition of low-dose steroids to intravenous gamma-globulin. Am J Obstet Gynecol 1996;174:1414–23.
13. Overton TG, Duncan KR, Jolly M, Letsky E, Fisk NM. Serial aggressive platelet transfusion for fetal alloimmune thrombocytopenia: platelet dynamics and perinatal outcome. Am J Obstet Gynecol 2002;186:826–31.
14. Paidas MJ, Berkowitz RL, Lynch L, Lockwood CJ, Lapinski R, McFarland JG, et al. Alloiummune thrombocytopenia: Fetal and neonatal losses related to cordocentesis. Am J Obstet Gynecol 1995;172:475–9.
15. Gaddipati S, Berkowitz RL, Lembet AA, Lapinski R, McFarland JG, Bussel JB, et al. Initial fetal platelet counts predict the response to intravenous gammaglobulin therapy in fetuses that are affected by PLA1 incompatibility. Am J Obstet Gynecol 2001;185:976–80.