Berkowitz, Richard L. MD1; Kolb, E Anders MD2; McFarland, Janice G. MD3; Wissert, Megan RN4; Primani, Andrea4; Lesser, Martin5; Bussel, James B. MD4
Fetal and neonatal alloimmune thrombocytopenia is a result of parental incompatibility of a platelet-specific antigen, most frequently PIA1 (HPA-1a), which affects approximately 1 in 1,000 live births.1 This disorder is the most common cause of severe thrombocytopenia in fetuses and term neonates2 and the most frequent cause of intracranial hemorrhage in term newborns.3 Ten percent to 22% of infants severely affected by alloimmune thrombocytopenia have an intracranial hemorrhage, and as many as 75% of these bleeds occur antenatally.2,4,5
Because population screening for alloimmune thrombocytopenia is not routinely performed, couples are usually identified to be at risk for this disorder after they have already had an affected infant. Presentation of the first affected child can range from incidentally detected neonatal thrombocytopenia to a devastating intracranial hemorrhage detected in utero. Therefore, testing for this disorder should be performed for any neonate with marked or unexplained thrombocytopenia.2
Once diagnosed, alloimmune thrombocytopenia should be treated antenatally because of its tendency to worsen in subsequent pregnancies.4 The goal of this treatment before birth is to eliminate both in utero and peripartum intracranial bleeding. Maternally administered intravenous immunoglobulin (IVIG) has been the most successful therapy explored to date in restoring adequate fetal platelets counts and preventing intracranial hemorrhage.6–12 Primary treatment with prednisone has not been as effective, but combining prednisone 1 mg/kg/d and IVIG 1 g/kg/wk substantially increases fetal platelet counts in fetuses who do not respond to IVIG alone.8
This report describes parallel, randomized, multicenter studies designed to evaluate the safety and efficacy of risk-based, stratified antenatal treatment for women who had previously delivered a fetus affected by alloimmune thrombocytopenia and were pregnant again with an affected fetus. Those patients whose previously affected fetus had suffered an antenatal intracranial hemorrhage were not included in this study, and the results of their care will be reported separately.
Between May 24, 1994, and April 12, 2001, 78 women with 79 separate pregnancies were enrolled at 42 institutions participating in a randomized treatment trial of fetuses affected by alloimmune thrombocytopenia (Table 1). It was necessary to recruit patients from all over the United States because of the relative rarity of the disease. 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 treatments of all patients was performed at the coordinating center and communicated to the participating institutions by telephone.
High-risk patients had either a sibling with an intracranial hemorrhage that occurred during the peripartum period or an initial fetal platelet count of less than 20,000/mL3. Treatment was initiated after fetal blood sampling at 20 weeks or later and was randomized between IVIG 1 g/kg/wk alone or in combination with prednisone 1 mg/kg/d. The standard-risk arm consisted of patients who did not have a sibling with an intracranial hemorrhage and in whom the initial platelet count in utero was greater than 20,000/mL3 but less than 100,000/mL3. Those women were randomly assigned to receive either IVIG 1 g/kg/wk or prednisone 0.5 mg/kg/d after fetal blood sampling as close to 20 weeks as possible.
Mothers and their fetuses were included in this series if a previous sibling had had alloimmune thrombocytopenia with or without a peripartum hemorrhage and if the current fetus was affected. The exclusion criteria were a prior affected fetus that had suffered a known antenatal intracranial hemorrhage or a maternal allergy to IVIG. A history of peripartum hemorrhage was defined as an intracranial hemorrhage discovered during the neonatal period and included cases that might have occurred during delivery.
Maternal–fetal PIA1 (HPA-1a) incompatibility, or that of another platelet specific antigen, was confirmed for all patients before enrollment by parental platelet antigen typing and maternal platelet antibody testing at the Blood Center of South Eastern Wisconsin.13 If the father was homozygous for the antigen in question, the fetus was presumed to be affected. If the father was heterozygous, the fetal genotype was determined by typing of amniocyte DNA.14 This protocol was performed under Investigation of New Drug (IND) application from the US Food and Drug Administration, approved by the institutional review boards of all participating institutions, and signed informed consent was obtained from all patients before enrollment.
All women in this series underwent an initial fetal blood sampling procedure before therapy was started. Subsequent blood sampling procedures were scheduled at 3- to 8-week intervals.6–8 The procedures were either performed by a local perinatologist or referred to New York City for performance by one of the authors (R.L.B.).
Under ultrasound guidance, an experienced operator inserted a 20- or 22-gauge spinal needle through the maternal anterior abdominal wall into an umbilical vessel. The mean corpuscular volume of aspirated red blood cells was used to confirm that the blood was fetal. It was recommended that, if the fetal platelet count was determined to be less than 50,000/mL3 when the needle was still in the sampled vessel, a transfusion of maternal platelets be administered.15 However, as some of these procedures were performed at multiple study sites, it is not certain that this was always done.
The protocol for assessment of response and intensification of therapy are delineated in Table 2. When fetal blood sampling was not possible for technical reasons, therapy was empirically intensified, and these patients were not included in the analysis of response to initial therapy. Whenever IVIG was administered at 2 g/kg/wk, it was infused in divided doses of 1 g/kg twice a week. Fetal ultrasound examinations to document growth and detect intracranial hemorrhage were performed every 2 to 4 weeks, and a neonatal intracranial ultrasound examination was obtained shortly after delivery.
Therapy was assigned for patients in the high-risk and standard-risk arms using a computer-generated random number list balanced by computed blocks. The Wilcoxon rank sum test was used to calculate the significance of differences in platelet counts and response to therapy and to compare the differences between birth platelet counts of neonates enrolled in the study and their untreated prior siblings. Fisher exact test was used to compare the responses in each treatment arm. Group-specific sample sizes could not be adequately powered to identify small differences among the arms of each study because of the rarity of the disorder. Therefore, for each of the parallel studies, there was only an 80% chance of detecting a 50–90% difference in response to therapy.
Alloimmune thrombocytopenia in 74 of the 79 pregnancies included in this study was caused by HPA-1a incompatibility. No fetus had a history of a previous sibling known to have an antenatal intracranial hemorrhage before delivery.
The 40 women in the high-risk arm either had a previous child with a peripartum intracranial hemorrhage (n = 7) and/or an initial fetal platelet count less than 20,000/mL3. Nineteen (48%) were randomly assigned to receive IVIG and prednisone, and 21 were given IVIG alone. The median initial fetal counts before therapy were 8,500/mL3 and 7,000/mL3, respectively. Five of the 7 fetuses whose previous siblings had a perinatal intracranial hemorrhage had initial platelet counts of < 10,000/mL3. Sixteen of 18 evaluable women (89%) receiving combination therapy had a satisfactory initial response compared with only 7 of 20 (35%) receiving IVIG alone (P < .05) (Table 2). Mean changes in fetal platelet counts between the first and second samplings were 67,100/mL3 and 17,300/mL3, respectively (P < .001). Twenty-two (55%) fetuses had initial fetal platelet counts less than 10,000/mL3. In that subset, 9 (82%) of 11 had a satisfactory response to treatment with IVIG and prednisone compared with only 2 (18%) of 11 treated with IVIG alone (P < .03).
The mean birth platelet counts, 99,400/mL3 and 66,800/mL3, respectively, were significantly higher than the mean initial neonatal platelet count of the 37 affected siblings from prior pregnancies in whom those data were available (24,000/mL3, P < .001 for each treatment group).
One fetus in the high-risk group developed an intracranial hemorrhage. This occurred in a patient treated with IVIG alone whose initial platelet count was 56,000/mL3 at 22 weeks of gestation. Because repeat fetal sampling at 28 weeks was not technically possible, treatment was empirically intensified by adding prednisone 1 mg/kg/d. The infant was delivered at 35 weeks of gestation with a birth platelet count of 14,000/mL3. A grade I intracranial hemorrhage was noted on the first day of life.
The 39 women in the standard-risk arm did not have a prior child who had suffered an intracranial hemorrhage, and the initial fetal platelet counts were all greater than 20,000/mL3. Nineteen were randomly assigned to receive IVIG 1 g/kg/wk, and 20 to prednisone 0.5 mg/kg/d. There were no significant differences in response to the treatments. Ten of 19 fetuses in women randomly assigned to IVIG had satisfactory initial platelet responses, compared with 12 of 19 evaluable fetuses on the prednisone arm. The mean fetal platelet increases in the first 3 to 8 weeks of treatment were 30,600/mL3 for IVIG and 25,700/mL3 for prednisone. Thirty-three neonates had birth platelet counts greater than 50,000/mL3, and overall values were significantly higher than the birth platelet counts of the 38 affected siblings from prior pregnancies (P < .001).
One patient on each arm of this group had an unexplained fetal demise. The first occurred at 32 weeks of gestation, 2 weeks after fetal blood sampling revealed the platelet count to be 96,000/mL3. The other occurred in a very unusual case in which the fetus was HPA-1a/1a but was being carried by an HPA-1b/lb surrogate.16 The fetal platelet count was 50,000/mL3 at 28 weeks of gestaion, and the fetal demise occurred 4 weeks later.
There were two intracranial hemorrhages in this group. The first was a grade 1 intracranial hemorrhage in a fetus born at 38 weeks of gestation with a birth platelet count of 172,000/mL3. The second was a grade 3 intracranial hemorrhage in an infant with a birth platelet count of 68,000/mL3, delivered at 28 weeks of gestation for persistent bradycardia after fetal blood sampling.
There were 11 serious complications after a total of 175 (6%) fetal blood sampling procedures. These included one fetal demise from unrecognized trauma 4 days after the procedure, 9 cesarean deliveries performed emergently at the time of the procedure, and 1 case of premature rupture of membranes 4 days after the procedure with an emergency delivery the following day for a nonreassuring fetal heart rate tracing. The 9 emergency cesarean deliveries were performed for persistent streaming in one case, prolonged bradycardia in seven, and persistent tachycardia followed by worrisome variable decelerations in the other. Nineteen (24%) infants in this series delivered before 34 weeks of gestation. Six of those deliveries occurred within 1 day of a fetal blood sampling procedure, and two from 1 to 7 days later. One of the infants was delivered at 24 weeks of gestation and died from complications of severe prematurity.
Patients whose fetus had an antenatal intracranial hemorrhage in a previous affected pregnancy were excluded from this study because the documented severity of their disease placed them in a unique category. The patients included in this report were arbitrarily separated into high-risk and standard-risk groups on the basis of a platelet count determined before therapy because it is known that fetuses with initial platelet counts less than 20,000/mL3 are less likely to respond to IVIG alone and are presumably at greater risk to develop an intracranial hemorrhage than those with higher counts.10,13 Among those women defined as being high risk, IVIG combined with prednisone was clearly more effective than IVIG alone in eliciting a satisfactory fetal platelet response. In the subgroup of patients with especially severe fetal thrombocytopenia at the time of initial sampling (ie, less than 10,000/mL3), IVIG 1 g/kg/wk plus prednisone 1 mg/kg/d satisfactorily increased the fetal platelet count in 82% of cases compared with 18% when IVIG was administered alone.
In contrast to the high-risk group described above, the patients with no history of intracranial hemorrhage in previous affected siblings and initial fetal platelet counts greater than 20,000/mL3 appear to benefit from less intensive therapy, regardless of type. Specifically, there was no significant advantage of IVIG alone over prednisone 0.5 mg/kg/d. Furthermore, only about 25% of patients from either arm required intensified therapy, and all but six neonates had birth platelet counts greater than 50,000/mL3. It must be noted, however, that this study was underpowered to demonstrate relatively small differences in these treatment arms.
In this and all our other studies6–8,10,13 to date, we have used serial fetal blood sampling procedures to initially assess the severity of fetal thrombocytopenia and then follow the effectiveness of the therapy being administered. The performance of percutaneous umbilical blood sampling, however, is clearly associated with an increased risk of fetal morbidity and mortality.15 As a consequence, some practitioners prefer to avoid performing that procedure and choose to treat alloimmune thrombocytopenia empirically,12 but there are drawbacks to this approach. Therapy with IVIG is expensive, and both IVIG and prednisone can cause adverse maternal side effects.8 This may be acceptable if the regimen being administered accomplishes its therapeutic objectives, but empiric therapy without knowing the fetal platelet count may be either unnecessary or inadequate. The former needlessly overtreats the mother while the latter allows the fetal platelet count to remain dangerously low. The precise role of this invasive procedure in the management of alloimmune thrombocytopenia remains to be determined.
In this study, we have documented that fetal platelet counts reflect a spectrum of disease severity and that fetuses with marked thrombocytopenia require more intensive therapy than those with higher initial levels of circulating platelets. The cost of obtaining that information, however, was substantial. Serious complications occurred in 6% of the fetal blood sampling procedures, and this resulted in the emergent delivery or death in utero of 11 (14%) of the 79 infants in this series. Birchall et al11 also identified a high rate of complications associated with sampling in patients with alloimmune thrombocytopenia.
A limitation of this study is the fact that, because of the rarity of alloimmune thrombocytopenia, the patients were treated in multiple sites around the United States by a group of perinatologists who could not be monitored directly. Perhaps if all of the fetal blood sampling procedures were performed by a handful of experts the results would have been better. This, however, is paradoxically one of the study's strengths because it reflects the reality of results obtained from well-intentioned, certified subspecialists working in different settings but adhering to a common research protocol.
Because fetal blood sampling is currently not frequently performed in the United States, it could be argued that optimal therapy might be obtained by referring all women with alloimmune thrombocytopenia to regional centers having proven expertise with that procedure. This, however, may not be practical. On the other hand, if one avoids performing percutaneous umbilical blood sampling because of concern about its associated fetal/neonatal morbidity, it follows that all of these patients must be treated empirically. In that case, what dose should be used? It seems reasonable to direct therapy to the most severely affected fetuses, not the least, for fear of undertreating a fetus with severe thrombocytopenia that goes on to have an intracranial hemorrhage. The results of the current investigation suggest that that dose should be IVIG 1 g/kg/wk plus prednisone 1 mg/kg/d starting at 20 weeks of gestation. This, however, will undoubtedly be associated with more adverse maternal side effects than if IVIG was administered alone.
In summary, this study demonstrates important biologic differences among fetuses with alloimmune thrombocytopenia as manifested by the relationship of their starting platelet count to their response to therapy. The standard-risk group appears to respond well to either IVIG 1 gm/kg/wk or prednisone 0.5 mg/kg/d. This is in marked contrast to the substantial number of patients in the high-risk group with an initial count of less than 10,000/mL3 for whom IVIG 1 gm/kg/wk was shown to be inadequate. The fetal–neonatal morbidity and mortality associated with fetal blood sampling was substantial, and this suggests that empiric therapy has the potential to be significantly safer for infants who meet the inclusion criteria for this study. If that were done, however, it should be recognized that therapy that maximizes platelet response in the most severely affected fetuses will overtreat others and certainly be associated with maternal side effects that could be avoided if daily high-dose prednisone was not administered. Finally, it must be emphasized that this study excluded patients whose fetus had an antenatal intracranial hemorrhage in a prior pregnancy. Treatment and outcome in that group will be addressed in a subsequent publication.
1. Williamson LM, Hackett G, Rennie J, Palmer CR, Maciver C, Hadfield R, et al. The natural history of fetomaternal alloimmunization to the platelet-specific antigen HPA-1a (PIA1, Zwa) as determined by antenatal screening. Blood 1998;92:2280–7.
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. Chaoying M, Junwu G, Chituwo BM. Intraventricular haemorrhage and its prognosis, prevention and treatment in term infants. J Trop Pediatr 1999;45:237–40.
4. Bussel JB, Zabusky ME, Berkowitz RL, McFarland JG. Fetal alloimmune thrombocytopenia. N Engl J Med 1997;337:22–6.
5. Herman JH, Jumbelic MI, Ancona RJ, Kickler TS. In utero cerebral hemorrhage in alloimmune thrombocytopenia. Am J Pediatr Hematol Oncol 1986;8:312–7.
6. Bussel JB, Berkowitz RL, McFarland JG, Lynch L, Chitkara U. Antenatal treatment of neonatal alloimmune thrombocytopenia. N Engl J Med 1988;319:1374–8.
7. Lynch L, Bussel JB, McFarland JG, Chitkara U, Berkowitz RL. Antenatal treatment of alloimmune thrombocytopenia. Obstet Gynecol 1992;80:67–71.
8. Bussel JB, Berkowitz RL, Lynch L, Lesser ML, Paidas MJ, Kanhai HH, et al. Antenatal management of alloimmune thrombocytopenia: a randomized trial in fifty-five maternal-fetal pairs. Am J Obstet Gynecol 1996;174:1414–23.
9. Wenstrom KD, Weiner CP, Williamson RA. Antenatal treatment of fetal alloimmune thrombocytopenia. Obstet Gynecol 1992;80:433–5.
10. Gaddipati S, Berkowitz RL, Lembet AA, Lapinski R, McFarland JG, Bussel JB. Initial fetal platelet counts predict the response to intravenous gammaglobulin therapy in fetuses affected by PLA-1 incompatibility. Am J Obstet Gynecol 2001;185:976–80.
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. Radder CM, Brand A, Kanhai HH. A less invasive treatment strategy to prevent intracranial hemorrhage in fetal and neonatal alloimmune thrombocytopenia. Am J Obstet Gynecol 2001;185:683–8.
13. Bussel JB, Kaplan C, McFarland J, The Working party on Neonatal Immune Thrombocytopenia of the Neonatal Hemostasis Subcommittee of the Scientific and Standardization Committee of the International Society of Thrombosis and Hemostasis. Recommendation for the evaluation and treatment of neonatal alloimmune and alloimmune thrombocytopenia. Thromb Hemost 1991;65:631–4.
14. McFarland JG, Aster RH, Bussel JB, Gianopoulos JG, Derbes RS, Newman PJ. Prenatal diagnosis of neonatal alloimmune thrombocytopenia using allele-specific oligonucleotide probes. Blood 1991;78:2276–82.
15. Paidas MJ, Berkowitz RL, Lynch L, Lockwood CJ, Lapinski R, McFarland JG, et al. Alloimmune thrombocytopenia: fetal and neonatal losses related to cordocentesis. Am J Obstet Gynecol 1995;172:475–9.
16. Curtis BR, Bussel JB, Manco-Johnson MJ, Aster RH, McFarland JG. Fetal and neonatal alloimmune thrombocytopenia in pregnancies involving in vitro fertilization: a report of four cases. Am J Obstet Gynecol 2005;192:543–47.
Figure. No caption available.