British /ournu/ of Haernatology. 199 1 , 78, 42 5-429
ADONIS
0007104891001 59C
HPA-5b (Bra) neonatal alloimmune thrombocytopenia: clinical and immunological analysis of 39 cases c. K A P L A N , M. c. MOREL-KOPP,H.K R O L L , *v. KIEFEL,* N. SCHLEGEL,t N. C H E S N E L AND C. MUELLER-ECKHARDT* Znstitut National de Transfusion Sanguine, Service d'lmmunologie Leuco-Plaquettaire, Paris, France, *Institute for Clinical Immunology and Transfusion Medicine, University of Giessen, Germany, and TH6pital Robert Debre, Luboratoire d'Hematologie et d'lmmunologie Clinique, Paris, France
Received 9 November 1990; uccepted for publication 12 March I991
Summary. Maternal alloimmunization against fetal platelets can cause fetal and neonatal thrombocytopenia (NAIT). The HPA-la (PI"', ZW') antigen is by far the most common antigen implicated in NAIT. However, today another antigen often linked with that affection is HPA-5b (Br"). This is a report of 39 cases of NAIT involving the HPA-5b antigen. Thrombocytopenia may be of grave consequence. Three infants developed intracerebral haemorrhages (ICH). Of these, one died presumably as a consequence of ICH. Central nervous system (CNS) sequelae in the neonatal period was
observed in two children. The potential hazards of death or disabling neurologic sequelae following intracerebral haemorrhage call for rapid and reliable diagnosis and effective therapy. Because there is high risk that subsequent pregnancies might be also affected by NAIT, the mothers of a previously affected child should be managed similarly to the HPA-lb mothers (PIA2.Zwh). The antenatal diagnosis of thrombocytopenia should be made and if necessary the in utero therapy instituted.
Neonatal alloimmune thrombocytopenia (NAIT)results from the maternal immunization against fetal platelet antigens (Harrington et a/. 1953: Shulman et a/. 1964). NAIT is considered to be the counterpart of the haemolytic disease of the newborn (HDN). However. in contrast to HDN. NAIT may already affect the first child. The potential hazards ofdeath or disabling neurological sequelae following haemorrhage of the central nervous system (CNS) call for rapid and reliable diagnosis and effective therapy. NAIT is due to the presence of alloantibodies directed against platelet specific alloantigens. Controversy exists whether HLA-antibodies. which are frequently encountered in sera of pregnant women, might induce this rather rare condition. A prospective analysis of 1507 sera from pregnant women suggests that HLA antibodies are not capable of causing thrombocytopenia in the newborn (Sharon & Amar. 198 1 ). It is possible that in those rare cases where neonatal thrombocytopenia was attributed to HLA antibodies, additional platelet specific alloantibodies were not diagnosed serologically or that other causes, not related to fetomaternal immunization, accounted for the thrombocytopenia. Casual observations suggest that NAIT is sometimes due to ABO incompatibility, although the particular features of these cases are not well defined.
Currently, five platelet alloantigen systems have been established* (Von dem Borne & Kuijpers, 1990). HPA-la (PI"'. Zw") is the antigen most frequently involved in alloimmunization of populations of European extraction. Recently the HPA-5 (Br)diallelic alloantigen system has been described (Kiefel et a/. 1988. 1989) using the newly developed monoclonal antibody-mediated immobilization of platelet antigen (MAIPA) assay (Kiefel et d. 1987). The allelic antigens Br"/Brh are inherited in a n autosomal codominant mode. Phenotype frequencies are 20-50/,for the Br" antigen in the German and 23% in the French population (Kiefel et a/. 1989; Bierling et a/, 1989) and 99.2% for the Brh antigen in Germany (Kiefel et a/. 1989). The Br antibodies are known to react weakly in conventional serological techniques using intact platelets such as immunofluorescence or radioimmunoassays. This is due to the low number of binding sites on platelets of Br"/Br" homozygous individuals (about 2000). Antibodies against the Br-antigens are not detected by immunoblotting which indicates that the respective alloantigens on glycoprotein Ia are combinatorial at sites sensitive to reduction. Preliminary data obtained from a rather small
Correspondence: Dr Cecile Kaplan. I.N.T.S.. h rue A. Cabanel. 7501 5 Paris. France.
5: Br. The allele of high frequency is called 'a', of low frequency 'b' (British ]ournu/ of Haernatology. 74. 2 39-240. 1990).
* The new nomenclature of platelet-specific alloantigens is: HPAI: Zw. PI": HPA-2:KO.Sib: HPA-3: Bak. Lek: HPA-4:Pen, Yuk: HPA-
425
426
C. Kaplan et al
cohort of NAIT cases caused by BP incompatibility appear to indicate that the bleeding tendency in these cases is milder than in cases with Zw" incompatibility (Kiefel et al. 1989: Mueller-Eckhardt et al. 1989a). In this study we report clinical and serological data on 39 BP-NAIT cases, emphasizing that when this affection is severe it can be associated with a considerable risk of intracerebral haemorrhage (ICH). PATIENTS AND METHODS The diagnosis of NAIT in all cases was based on the typical clinical course of the newborn and the presence of Br" antibodies in maternal sera. Cases were included in this study only when the platelets of the fathers and, if tested, of the children were BP positive. Serology. Platelets were typed with the platelet indirect radioactive assays for Zw (Soulier ~t al, 1975), and with the MAIPA test (Kiefel et al. 1987) for Bak and Br. Maternal sera were screened with the same techniques for platelet reactive antibodies. HLA antibodies were determined by microlymphocytotoxicity (Terasaki et al. 1978). Platelet panels were collected from healthy donors tested for Zw". Zwb, BP, Brb, Baku and Bak". Clinical evaluation. This study was in part retrospective. Patients with clinical diagnosis of NAIT. but with negative or equivocal serological results tested before 198 7 were reinvestigated when the MAIPA test became available. In three families, two successive pregnancies were followed.
Table I. Haemorrhagic symptoms in children with Br'-NAIT
as compared to those with Zw"-NAIT
None Purpura/haematoma Visceral haemorrhage
Br"
ZW"
( n = 39)
( n = 127)*
2 3 (59%)
1 7 ( 1 3%) 8 3 (65%) 7 (6%) 2 0 ( 1 6%)
7 (18%) 6 (15%)
ICH
3 (8%)
* Kaplan et a1 (1990).
Table 11. Distribution of children with platelet counts in categories
< l o . 10 3 0 ( x 10"/1) born to mothers with anti Zw" and anti-Br", respectively Nadir
Postpartum
Platelet counts < 10 10-30 > 30
a-Zwa*
a-BrU
a-Zwa8
a-Br"
22
1
38
4
38 21
6 32
39
13
4
22
x2= 34.22 (DF=2)
RESULTS Serological evaluation Bra antibodies were present in all cases. They were the only antibodies found in 12 cases, whereas they were associated with anti-Zw" in two instances and with anti-Bak" in one instance. In addition, HLA antibodies were present in 2 5 cases. No information was available concerning HLA antibodies for two sera. One mother typed Zw" negative and BP negative. was incompatible with her child for both antigens, but her serum reacted only with Br" positive platelets. Clinical rvaluation Only 12 infants (30%)were born to primiparous mothers. More often the mother was multiparous and NAIT was diagnosed after three or more pregnancies (21 cases). Abortions were reported in 14 women. The duration of pregnancy was normal except in six cases. The birth weight was within the normal range in 30 of the 39 neonates studied. 58% of the deliveries were vaginal. The remainder 42%. because of fetal distress, were caesarean sections. The sex ratio of children was 2.9: 1 (29 boys, 10 girls). At birth. 59%of the children did not exhibit haemorrhagic symptoms. Petechiae were found in seven cases. In nine cases visceral haemorrhage and/or ICH were noted (Table I). Other findings included icterus in eight cases, hepatomegaly or hepatosplenomegaly in four, Down's syndrome in one, Pierre Kobin syndrome in one, infections in five. Cytomegalovirus (CMV) infection was diagnosed in two infants. Another child developed streptococcus-D infection. The remaining two had
P< lo-'
The differencesare highly signillcant (xz test for 2 x 3 contingency tables). Mean (m) and range of platelet counts: anti-Zw? post-partum m=28*4, range 1-160; nadir m=16.9. range 1-150. anti-BP: post-partumm=64.2, range 2-140 nadum=44.6. range 2-123. * Mueller-Eckhardt et al(l989a).
presumptive diagnosis of infection based on the existence of maternal fever a t the time of delivery. Four infants developed anaemia following significant haemorrhage. Platelet counts at birth higher than 30 x 10y/l declined further in the post-partum period in 23 out of 32 cases. The nadir was usually reached at day 3 (mean 3.4 d. range 0 - 1 1 ) (Table 11). Bone-marrow aspirates were performed in two cases and showed few or no megakaryocytes. In one of them the outcome was good while the other followed a very severe course (case 1 ) . Therapy varied according to the severity of symptoms (Table 111). Transfusion of maternal platelets immediately increased the infants' platelet counts. Exchangetransfusion or random platelet transfusions had no or only a marginal effect in some cases and did not always produce a significant and sustained increase of platelet count. Nine neonates were treated with intravenous immunoglobulins (IVlgG). Five of them received additional therapeutic regimens. All four children who had received only IVlgG responded favourably, while the effect of the IVIgG in the
HPA- 5b Neonatal Alloimmune Thrombocytopeniu Table Ill. Therapy
No. patients None
20
Single IVIgC Random platelet transfusion Exchange transfusion
4 6
Combined Exchange transfusion followed by maternal transfusion Random platelet transfusion followed by IVlgC and/or corticosteroids
2
2
5
remainder five children could not be assessed. The clinical outcome was favourable in 30 children. Platelet counts below 100 x 10y/l lasted for 10 d on average (range 2-32 d). One child died of ICH. while neurological sequelae were observed in two others. The more severe cases of our series due to Br" antibodies alone are briefly described below. Case reports Case 1. This boy was delivered by a IV parous. IV gravida mother at week 38 of gestation by caesarean section due to pre-eclampsia and fetal distress. Immediately after birth the neonate had to be treated for asphyxia. At that time his platelet count was 44 x 1OY/l. There were signs of tracheal haemorrhages. At day 5 the platelet count fell to 8 x 10y/l and a head sonogram showed signs of ICH. Despite random platelet transfusions (days 5 and 9 ) and an exchange transfusion on day 9 before the second platelet transfusion. the general child's condition deteriorated with seizures appearing on day 9. On the following day, a grade IV ICH was diagnosed by sonogram and CT scan: the baby died on day 10. Case 2. This case concerns a seventh child born to a 41year-old woman. No signs of NAIT had been documented in the previous children. The present pregnancy was uneventful with a spontaneous delivery at term. Immediately post partum petechiae were noted. The platelet count was 25 x 10q/I and haemoglobin was 6 g/dl. Haematemesis occurred a few hours later with thrombocytopenia of 1 0 x 10y/l.There were no signs of ICH. The infant received 1 unit of whole blood and a random platelet transfusion. The platelet count increased to 50 x 10y/land was 1 3 5 x loy/]at day 20 upon discharge. CUSP 3. This male infant was delivered at term by a caesarean section performed for fetal bradycardia and meconium-stained fluid. His mother was 29 years old. 111 parous. IV gravida. At birth the infant had to be treated for asphyxia. 1 0 min later signs of haemorrhage appeared: purpura. haemoptyses. haematemesis. haematuria. The platelet count was 3 0 x 1Oy/l and fell to 1 0 x 1 Oy/lin spite of two exchange
42 7
transfusions. Head sonograms were normal and no ICH occurred. Thrombocytopenia persisted for 1 month and then slowly disappeared. The child was discharged in good condition. Case 4. This child, sibling of case 3. was born at term by caesarean section. At birth petechial purpura was seen with haematomas. The platelet count was 52 x I OY/1 associated with mild anaemia (haemoglobin 11 g/dl). On day 2. haemorrhage was treated by an exchange transfusion. Because of severe thrombocytopenia (nadir 20 x 1Oy/l) maternal platelets were given with an immediate positive response (post-transfusion platelet count 395 x 1 OY/I). The further course was uneventful. The head CT scan at discharge was normal. Case 5. During the 34th and 36th weeks of pregnancy of this V parous V gravida mother, sonogram examinations revealed a cerebral ventricular dilatation of the fetus. Pregnancy was interrupted at 39.5 weeks by a caesarean section because of fetal distress, bradycardia, meconium-stained fluid and ventricular dilatation. After delivery the infant exhibited petechiae and hepatosplenomegaly. Head sonogram revealed ICH. Because of positive CMV antibody tests, a cytomegalovirus infection was suspected. The fetal platelet count was 10 x 10y/lprompting a n immediate random platelet transfusion. The platelet count increased to I 0 0 x 1 OY/1 but fell to 25 x 1OY/1 1 d later. On day 11 the platelet count reached 56 x 1 Oy/l and was almost normal ( 12 3 x 1Oy/l) at day 14. A repeat sonogram a t day 20 revealed residual ICH. The clinical course was marked by neurological sequelae. Case 6. A male infant with symptoms of Pierre-Robin syndrome was delivered spontaneously by a I1 parous. I1 gravida mother. 6 h post partum the infant exhibited petechiae and haematemesis. The platelet count was 3 1 x 10y/l. He received daily intravenous gammaglobulins for three consecutive days (days 2-5) in conjunction with corticosteroids and a random platelet concentrate on day 4. The platelet count normalized on day 1 3 . The outcome was good. Case 7. This premature ( 33 weeks of gestation) male infant was born to a IV gravida, I parous mother by a caesarean section because of fetal distress. At birth, he exhibited generalized petechiae with a platelet count of 80 x 1 0 y / l . At day 3 he received packed red blood cells as treatment for haematemesis and intra-pulmonary haemorrhage. Subsequently he was prescribed intravenous gammaglobulins 0.4 g/kg for 5 d in conjunction with corticosteroids. The platelet count increased to 200 x 1OY/1. A severe icterus appeared. CMV infection was diagnosed. The outcome was favourable with absence of CNS lesions. lniniunobgical studies Eleven out of 1 5 (68.8%)Caucasoid mothers were found to be HLA DRw6 (control: 28.7%:n=408).
DISCUSS 1O N Fetomaternal platelet incompatibility due to anti-Bra. first described in 1988 by Kiefel er a/. was soon found to be the second most frequent cause of NAIT in the German and
428
C. Kaplan et a1
French populations. In our recent survey (Mueller-Eckhardt rt al, 1989a). 19% of NAIT cases elicited by serologically verified platelet specific alloantibodies were due to anti-BP. Similar data were found in the present material from Paris: 60% of definitively identified NAIT cases were due to a Zw", 26% to a BP incompatibility. Most recent information from Japan seems to indicate that BP antibodies rank first as the cause of NAIT in the Japanese population (Y. Shibata, personal communication). Nevertheless. precise 5gures as to the relative frequency of BP-NAIT in the various ethnic groups are not available. Our study presents 39 cases with BP-NAIT. and reports that the bleeding tendency in this particular condition is milder than in Zw"-NAIT. Thrombocytopenia in children was usually mild, 23 out of 39 infants (59%) lacked haemorrhagic symptoms and 2 0 children did not require any treatment. This sharply contrasts to the low number (13%) of asymptomatic neonates afflicted by Zw" incompatibility (Table I). Notwithstanding, severe haemorrhages were occasionally observed. Six of our cases (15%) suffered visceral haemorrhages. Three had ICH. Of these three, one child died as consequence of the severity of the ICH. Another was left with no neurological sequelae. While the remaining child suffered also an infection by CMV. This superimposed infection makes it difficult to link all the CNS sequelae only to ICH (case 5). In our series, another child was found to have CNS sequelae: however, no data exists to allow us to correlate these findings with the presence of ICH at birth. Our data provide evidence to support that the incidence of neurological complications in BP-NAIT is lower than in Zw"-NAIT. However, a larger series will be necessary before one can draw definite conclusions. Therapy of BP-NAIT consists, as in all fetomaternal incompatibilities. in providing antigen-negative platelets, either from the mothers or from phenotyped donors. As expected, the two children in our series who received maternal platelets had a positive and sustained response. The response of 12 neonates receiving random platelets was not clear. Because of the phenotype distribution of BP in the central European populations (about 20%) it is to be anticipated that platelets of four out of five random donors should be compatible. Accordingly, we have seen good responses in a number of cases treated by exchange transfusion and/or random platelets. In cases where repetitive random platelet transfusions were ineffective, other causes for refractoriness (i.e. sepsis, hepatosplenomegaly, insufficient platelet preparation) cannot definitively be excluded. As it has been shown for Zw"-NAIT (Mueller-Eckhardt et al. 1989a), IVIgG is a good therapeutic alternative. The effect of IVIgG could only be properly evaluated in four children in our series. These cases responded favourably. The mode of immunization in BP incompatibility appears somewhat at variance from that established for Zwa immunization. While in the latter condition approximately 50% of the mothers develop Zw" antibodies against fetal antigens already during the first pregnancy, BP antibodies were often detected for the first time in multiparous women in our series. Whether this late antibody recognition reflects 'boostered' sensitization of the mother by successive pregnancies or the
condition did exist previously but only caused mild thrombocytopenia in the previous children, and thus went undiagnosed, remains to be determined. We hypothesize that BP immunization probably arises as early as Zw" immunization, particularly in view of the fact that 'responders' to Zw" and BP immunization have a similar immunogenetic background strong association with HLADR3, DRw52 (Reznikoff-Etievant et al. 1983. 1988: Valentin et al. 1990) or DRw6 (Mueller-Eckhardt et al. 1989b). respectively. We believe that in future prospective studies using sensitive techniques for BP antibody detection, i.e. the MAIPA assay, not only many more 'responders' will be found, but also that the prevalence of asymptomatic Br'NAIT will increase. Because the recurrence of BP-NAIT in families is likely to occur, management of mothers of previously affected children should follow the same rules as for Zw"-NAIT (Kaplan et al. 1988. 1990: Bussel, 1990). We believe that antenatal diagnosis of fetal thrombocytopenia by umbilical cord blood sampling should be recommended in all BP-NAIT mothers. However, we think that interventions should be reserved to cases with a history of severely affected previous children. Concise diagnostic and therapeutic guidelines, however, have yet to be outlined. ACKNOWLEDGMENTS We are grateful to the physicians who provided clinical information concerning their patients. We also thank M. C. Mateos and J. Quesne for the technical assistance and T. Caetano for secretarial assistance. This work was supported in part by Procope 1989-1 990. by the Fondation de France, and by the Deutsche Forschungsgemeinschaft (DFG) (Mu 277/9-7). REFERENCES Bierling, P.. Fromont. P.. Bettaieb. A. & Duedari. N. (1989) Anti-Br" antibodiesin the French population. British lournal of Haematology. 73,428-429.
Bussel, J.B. (1990) Antenatal treatment of neonatal alloimmune thrombocytopenia with intravenous gammaglobulin. Platelet Irnmuno1ogy:fundamental and clinical aspects (ed. by C. Kaplan-Gouet. N. Schlegel. Ch. Salmon and J. McCregor), pp. 287-295. John Libbey Eurotext. Paris. Harrington. W.J.. Sprague. C.C.. Minnich. V., Moore. C.V.. Aulvin. R.C. & Dubach. R. (1953) Immunologic mechanisms in idiopathic and neonatal thrombocytopenicpurpura. Annals of Jnternal Medicine, 38,433-469.
Kaplan. C., Daffos, F., Forestier. F.. Cox, W.L.. Lyon-Caen. D.,DupuyMontbrun. M.C. & Salmon, Ch. (1988) Management of alloimmune thrombocytopenia:antenatal diagnosis and in utero transfusion of maternal platelets. Blood. 72. 340-343. Kaplan C.. Ddos. P.. Forestier, F., Morel, M.C.. Chesnel. N. & Tchemia. G. (1990) Current trends in neonatal alloimmune thrombocytopenia: diagnosis and therapy. Platelet Immunology: fundamental and clinical aspects (ed. by C. Kaplan-Gouet, N. Schlegel, Ch. Salmon and J. McCregor).pp. 267-278. John Libbey Eurotext. Paris. Kiefel. V., Santoso. S.. Weisheit. M. & Mueller-Eckhardt, C. (1987) Monoclonal antibody-specific immobilization of platelet antigens
HPA-Sb Neonatal Alloimmune Thrombocytopenia (MAIPA): a new tool for the identification of platelet-reactive antibodies. Blood. 70, 1722-1 726. Kiefel. V.. Santoso. S.. Katzmann. B. & Mueller-Eckhardt. C. ( 1 988) A new platelet-specific alloantigen Bra. Vow Sanguinis. 54, 101-1 Oh. Kiefel, V.. Santoso. S.. Katzmann. B. & Mueller-Eckhardt. C. (1989) The BP/Brb alloantigen system on human platelets. Blood, 73, 2219-2223. Mueller-Eckhardt, C.. Kiefel. V.. Grubert. A.. Kroll, H.. Weisheit. M.. Schmidt. S.. Mueller-Eckhardt.G. & Santoso, S . ( 1 989a) 348 cases of suspected neonatal alloimmune thrombocytopenia. Lanret, i, 363-366. Mueller-Eckhardt. C., Kiefel, V.. Kroll. H. & Mueller-Eckhardt. G. (1989b) HLA-DRWh. a new immune response marker for immunization agalnst the platelet alloantigen BP. Vox Sanguinis. 57. 90-9 1 . ReznikofT-Etievant. M.F.. Muller. J.Y.. Julien. F. & Patereau. C. (1983) An immune response gene linked to HLA in man. Tissue Antigens. 22. 312-314. ReznikofT-Etievant, M.F.. Kaplan. C.. Muller. J.Y..DafTos. F. & Forestier. F. ( 1988) Alloimmune thrombocytopenias. definition of a group at risk: a prospective study. Current Studies in Hematology arid Blood Transfusion. 55, 119-124.
429
Sharon, R. & Amar. A. (1981) Maternal anti-HLA antibodies and neonatal thrombocytopenia. Lancet. i, 31 3. Shulman. N.R.. Marder. V.J., Hiller, M.C. & Collier, M. (1 964)Platelet and leucocyte iso-antigens and their antibodies: serologic, physiologic and clinical studies. Progress in Hematology (ed. by C. V. Moore and S. B. Brown). pp. 222-304. Grune and Stratton. New York. Soulier. J.P., Patereau. C. & Drouet, J. (1975) Platelet indirect radioactive Coombs test. Its utilisation for PL*’ grouping. Vox Sanguinis, 29, 253-268. Terasaki. P.I.. Bernoco, D.. Park, M.S..Oiturk. G. & Iwaki, Y. ( 1 978) Microdroplet testing for HLA A. B. C and D antigens. American lournal of Clinical Pathology. 69. 1 0 3-1 20. Valentin, N.. Vergracht. A.. Bignon. J.D.. Cheneau. M.L.. Blanchard. D.. Kaplan, C.. Reznikoff-Etievant. M.F. & Muller. J.Y. (1990) HLADRw52a is involved in alloimmunization against PLAi antigen. Human Immunology, 27, 73-79. Von dem Borne. A.E.G.Kr. & Kuijpers. R.W.A.M. (1990) Platelet antigens. new aspects. Platelet Zmmunology:fundamental and clinical aspects (ed.by C. Kaplan-Gouet, N. Schlegel, Ch. Salmon and J. McGregor). pp. 21 9-240. John Libbey Eurotext. Paris.
ADONIS
0007104891001 59C
HPA-5b (Bra) neonatal alloimmune thrombocytopenia: clinical and immunological analysis of 39 cases c. K A P L A N , M. c. MOREL-KOPP,H.K R O L L , *v. KIEFEL,* N. SCHLEGEL,t N. C H E S N E L AND C. MUELLER-ECKHARDT* Znstitut National de Transfusion Sanguine, Service d'lmmunologie Leuco-Plaquettaire, Paris, France, *Institute for Clinical Immunology and Transfusion Medicine, University of Giessen, Germany, and TH6pital Robert Debre, Luboratoire d'Hematologie et d'lmmunologie Clinique, Paris, France
Received 9 November 1990; uccepted for publication 12 March I991
Summary. Maternal alloimmunization against fetal platelets can cause fetal and neonatal thrombocytopenia (NAIT). The HPA-la (PI"', ZW') antigen is by far the most common antigen implicated in NAIT. However, today another antigen often linked with that affection is HPA-5b (Br"). This is a report of 39 cases of NAIT involving the HPA-5b antigen. Thrombocytopenia may be of grave consequence. Three infants developed intracerebral haemorrhages (ICH). Of these, one died presumably as a consequence of ICH. Central nervous system (CNS) sequelae in the neonatal period was
observed in two children. The potential hazards of death or disabling neurologic sequelae following intracerebral haemorrhage call for rapid and reliable diagnosis and effective therapy. Because there is high risk that subsequent pregnancies might be also affected by NAIT, the mothers of a previously affected child should be managed similarly to the HPA-lb mothers (PIA2.Zwh). The antenatal diagnosis of thrombocytopenia should be made and if necessary the in utero therapy instituted.
Neonatal alloimmune thrombocytopenia (NAIT)results from the maternal immunization against fetal platelet antigens (Harrington et a/. 1953: Shulman et a/. 1964). NAIT is considered to be the counterpart of the haemolytic disease of the newborn (HDN). However. in contrast to HDN. NAIT may already affect the first child. The potential hazards ofdeath or disabling neurological sequelae following haemorrhage of the central nervous system (CNS) call for rapid and reliable diagnosis and effective therapy. NAIT is due to the presence of alloantibodies directed against platelet specific alloantigens. Controversy exists whether HLA-antibodies. which are frequently encountered in sera of pregnant women, might induce this rather rare condition. A prospective analysis of 1507 sera from pregnant women suggests that HLA antibodies are not capable of causing thrombocytopenia in the newborn (Sharon & Amar. 198 1 ). It is possible that in those rare cases where neonatal thrombocytopenia was attributed to HLA antibodies, additional platelet specific alloantibodies were not diagnosed serologically or that other causes, not related to fetomaternal immunization, accounted for the thrombocytopenia. Casual observations suggest that NAIT is sometimes due to ABO incompatibility, although the particular features of these cases are not well defined.
Currently, five platelet alloantigen systems have been established* (Von dem Borne & Kuijpers, 1990). HPA-la (PI"'. Zw") is the antigen most frequently involved in alloimmunization of populations of European extraction. Recently the HPA-5 (Br)diallelic alloantigen system has been described (Kiefel et a/. 1988. 1989) using the newly developed monoclonal antibody-mediated immobilization of platelet antigen (MAIPA) assay (Kiefel et d. 1987). The allelic antigens Br"/Brh are inherited in a n autosomal codominant mode. Phenotype frequencies are 20-50/,for the Br" antigen in the German and 23% in the French population (Kiefel et a/. 1989; Bierling et a/, 1989) and 99.2% for the Brh antigen in Germany (Kiefel et a/. 1989). The Br antibodies are known to react weakly in conventional serological techniques using intact platelets such as immunofluorescence or radioimmunoassays. This is due to the low number of binding sites on platelets of Br"/Br" homozygous individuals (about 2000). Antibodies against the Br-antigens are not detected by immunoblotting which indicates that the respective alloantigens on glycoprotein Ia are combinatorial at sites sensitive to reduction. Preliminary data obtained from a rather small
Correspondence: Dr Cecile Kaplan. I.N.T.S.. h rue A. Cabanel. 7501 5 Paris. France.
5: Br. The allele of high frequency is called 'a', of low frequency 'b' (British ]ournu/ of Haernatology. 74. 2 39-240. 1990).
* The new nomenclature of platelet-specific alloantigens is: HPAI: Zw. PI": HPA-2:KO.Sib: HPA-3: Bak. Lek: HPA-4:Pen, Yuk: HPA-
425
426
C. Kaplan et al
cohort of NAIT cases caused by BP incompatibility appear to indicate that the bleeding tendency in these cases is milder than in cases with Zw" incompatibility (Kiefel et al. 1989: Mueller-Eckhardt et al. 1989a). In this study we report clinical and serological data on 39 BP-NAIT cases, emphasizing that when this affection is severe it can be associated with a considerable risk of intracerebral haemorrhage (ICH). PATIENTS AND METHODS The diagnosis of NAIT in all cases was based on the typical clinical course of the newborn and the presence of Br" antibodies in maternal sera. Cases were included in this study only when the platelets of the fathers and, if tested, of the children were BP positive. Serology. Platelets were typed with the platelet indirect radioactive assays for Zw (Soulier ~t al, 1975), and with the MAIPA test (Kiefel et al. 1987) for Bak and Br. Maternal sera were screened with the same techniques for platelet reactive antibodies. HLA antibodies were determined by microlymphocytotoxicity (Terasaki et al. 1978). Platelet panels were collected from healthy donors tested for Zw". Zwb, BP, Brb, Baku and Bak". Clinical evaluation. This study was in part retrospective. Patients with clinical diagnosis of NAIT. but with negative or equivocal serological results tested before 198 7 were reinvestigated when the MAIPA test became available. In three families, two successive pregnancies were followed.
Table I. Haemorrhagic symptoms in children with Br'-NAIT
as compared to those with Zw"-NAIT
None Purpura/haematoma Visceral haemorrhage
Br"
ZW"
( n = 39)
( n = 127)*
2 3 (59%)
1 7 ( 1 3%) 8 3 (65%) 7 (6%) 2 0 ( 1 6%)
7 (18%) 6 (15%)
ICH
3 (8%)
* Kaplan et a1 (1990).
Table 11. Distribution of children with platelet counts in categories
< l o . 10 3 0 ( x 10"/1) born to mothers with anti Zw" and anti-Br", respectively Nadir
Postpartum
Platelet counts < 10 10-30 > 30
a-Zwa*
a-BrU
a-Zwa8
a-Br"
22
1
38
4
38 21
6 32
39
13
4
22
x2= 34.22 (DF=2)
RESULTS Serological evaluation Bra antibodies were present in all cases. They were the only antibodies found in 12 cases, whereas they were associated with anti-Zw" in two instances and with anti-Bak" in one instance. In addition, HLA antibodies were present in 2 5 cases. No information was available concerning HLA antibodies for two sera. One mother typed Zw" negative and BP negative. was incompatible with her child for both antigens, but her serum reacted only with Br" positive platelets. Clinical rvaluation Only 12 infants (30%)were born to primiparous mothers. More often the mother was multiparous and NAIT was diagnosed after three or more pregnancies (21 cases). Abortions were reported in 14 women. The duration of pregnancy was normal except in six cases. The birth weight was within the normal range in 30 of the 39 neonates studied. 58% of the deliveries were vaginal. The remainder 42%. because of fetal distress, were caesarean sections. The sex ratio of children was 2.9: 1 (29 boys, 10 girls). At birth. 59%of the children did not exhibit haemorrhagic symptoms. Petechiae were found in seven cases. In nine cases visceral haemorrhage and/or ICH were noted (Table I). Other findings included icterus in eight cases, hepatomegaly or hepatosplenomegaly in four, Down's syndrome in one, Pierre Kobin syndrome in one, infections in five. Cytomegalovirus (CMV) infection was diagnosed in two infants. Another child developed streptococcus-D infection. The remaining two had
P< lo-'
The differencesare highly signillcant (xz test for 2 x 3 contingency tables). Mean (m) and range of platelet counts: anti-Zw? post-partum m=28*4, range 1-160; nadir m=16.9. range 1-150. anti-BP: post-partumm=64.2, range 2-140 nadum=44.6. range 2-123. * Mueller-Eckhardt et al(l989a).
presumptive diagnosis of infection based on the existence of maternal fever a t the time of delivery. Four infants developed anaemia following significant haemorrhage. Platelet counts at birth higher than 30 x 10y/l declined further in the post-partum period in 23 out of 32 cases. The nadir was usually reached at day 3 (mean 3.4 d. range 0 - 1 1 ) (Table 11). Bone-marrow aspirates were performed in two cases and showed few or no megakaryocytes. In one of them the outcome was good while the other followed a very severe course (case 1 ) . Therapy varied according to the severity of symptoms (Table 111). Transfusion of maternal platelets immediately increased the infants' platelet counts. Exchangetransfusion or random platelet transfusions had no or only a marginal effect in some cases and did not always produce a significant and sustained increase of platelet count. Nine neonates were treated with intravenous immunoglobulins (IVlgG). Five of them received additional therapeutic regimens. All four children who had received only IVlgG responded favourably, while the effect of the IVIgG in the
HPA- 5b Neonatal Alloimmune Thrombocytopeniu Table Ill. Therapy
No. patients None
20
Single IVIgC Random platelet transfusion Exchange transfusion
4 6
Combined Exchange transfusion followed by maternal transfusion Random platelet transfusion followed by IVlgC and/or corticosteroids
2
2
5
remainder five children could not be assessed. The clinical outcome was favourable in 30 children. Platelet counts below 100 x 10y/l lasted for 10 d on average (range 2-32 d). One child died of ICH. while neurological sequelae were observed in two others. The more severe cases of our series due to Br" antibodies alone are briefly described below. Case reports Case 1. This boy was delivered by a IV parous. IV gravida mother at week 38 of gestation by caesarean section due to pre-eclampsia and fetal distress. Immediately after birth the neonate had to be treated for asphyxia. At that time his platelet count was 44 x 1OY/l. There were signs of tracheal haemorrhages. At day 5 the platelet count fell to 8 x 10y/l and a head sonogram showed signs of ICH. Despite random platelet transfusions (days 5 and 9 ) and an exchange transfusion on day 9 before the second platelet transfusion. the general child's condition deteriorated with seizures appearing on day 9. On the following day, a grade IV ICH was diagnosed by sonogram and CT scan: the baby died on day 10. Case 2. This case concerns a seventh child born to a 41year-old woman. No signs of NAIT had been documented in the previous children. The present pregnancy was uneventful with a spontaneous delivery at term. Immediately post partum petechiae were noted. The platelet count was 25 x 10q/I and haemoglobin was 6 g/dl. Haematemesis occurred a few hours later with thrombocytopenia of 1 0 x 10y/l.There were no signs of ICH. The infant received 1 unit of whole blood and a random platelet transfusion. The platelet count increased to 50 x 10y/land was 1 3 5 x loy/]at day 20 upon discharge. CUSP 3. This male infant was delivered at term by a caesarean section performed for fetal bradycardia and meconium-stained fluid. His mother was 29 years old. 111 parous. IV gravida. At birth the infant had to be treated for asphyxia. 1 0 min later signs of haemorrhage appeared: purpura. haemoptyses. haematemesis. haematuria. The platelet count was 3 0 x 1Oy/l and fell to 1 0 x 1 Oy/lin spite of two exchange
42 7
transfusions. Head sonograms were normal and no ICH occurred. Thrombocytopenia persisted for 1 month and then slowly disappeared. The child was discharged in good condition. Case 4. This child, sibling of case 3. was born at term by caesarean section. At birth petechial purpura was seen with haematomas. The platelet count was 52 x I OY/1 associated with mild anaemia (haemoglobin 11 g/dl). On day 2. haemorrhage was treated by an exchange transfusion. Because of severe thrombocytopenia (nadir 20 x 1Oy/l) maternal platelets were given with an immediate positive response (post-transfusion platelet count 395 x 1 OY/I). The further course was uneventful. The head CT scan at discharge was normal. Case 5. During the 34th and 36th weeks of pregnancy of this V parous V gravida mother, sonogram examinations revealed a cerebral ventricular dilatation of the fetus. Pregnancy was interrupted at 39.5 weeks by a caesarean section because of fetal distress, bradycardia, meconium-stained fluid and ventricular dilatation. After delivery the infant exhibited petechiae and hepatosplenomegaly. Head sonogram revealed ICH. Because of positive CMV antibody tests, a cytomegalovirus infection was suspected. The fetal platelet count was 10 x 10y/lprompting a n immediate random platelet transfusion. The platelet count increased to I 0 0 x 1 OY/1 but fell to 25 x 1OY/1 1 d later. On day 11 the platelet count reached 56 x 1 Oy/l and was almost normal ( 12 3 x 1Oy/l) at day 14. A repeat sonogram a t day 20 revealed residual ICH. The clinical course was marked by neurological sequelae. Case 6. A male infant with symptoms of Pierre-Robin syndrome was delivered spontaneously by a I1 parous. I1 gravida mother. 6 h post partum the infant exhibited petechiae and haematemesis. The platelet count was 3 1 x 10y/l. He received daily intravenous gammaglobulins for three consecutive days (days 2-5) in conjunction with corticosteroids and a random platelet concentrate on day 4. The platelet count normalized on day 1 3 . The outcome was good. Case 7. This premature ( 33 weeks of gestation) male infant was born to a IV gravida, I parous mother by a caesarean section because of fetal distress. At birth, he exhibited generalized petechiae with a platelet count of 80 x 1 0 y / l . At day 3 he received packed red blood cells as treatment for haematemesis and intra-pulmonary haemorrhage. Subsequently he was prescribed intravenous gammaglobulins 0.4 g/kg for 5 d in conjunction with corticosteroids. The platelet count increased to 200 x 1OY/1. A severe icterus appeared. CMV infection was diagnosed. The outcome was favourable with absence of CNS lesions. lniniunobgical studies Eleven out of 1 5 (68.8%)Caucasoid mothers were found to be HLA DRw6 (control: 28.7%:n=408).
DISCUSS 1O N Fetomaternal platelet incompatibility due to anti-Bra. first described in 1988 by Kiefel er a/. was soon found to be the second most frequent cause of NAIT in the German and
428
C. Kaplan et a1
French populations. In our recent survey (Mueller-Eckhardt rt al, 1989a). 19% of NAIT cases elicited by serologically verified platelet specific alloantibodies were due to anti-BP. Similar data were found in the present material from Paris: 60% of definitively identified NAIT cases were due to a Zw", 26% to a BP incompatibility. Most recent information from Japan seems to indicate that BP antibodies rank first as the cause of NAIT in the Japanese population (Y. Shibata, personal communication). Nevertheless. precise 5gures as to the relative frequency of BP-NAIT in the various ethnic groups are not available. Our study presents 39 cases with BP-NAIT. and reports that the bleeding tendency in this particular condition is milder than in Zw"-NAIT. Thrombocytopenia in children was usually mild, 23 out of 39 infants (59%) lacked haemorrhagic symptoms and 2 0 children did not require any treatment. This sharply contrasts to the low number (13%) of asymptomatic neonates afflicted by Zw" incompatibility (Table I). Notwithstanding, severe haemorrhages were occasionally observed. Six of our cases (15%) suffered visceral haemorrhages. Three had ICH. Of these three, one child died as consequence of the severity of the ICH. Another was left with no neurological sequelae. While the remaining child suffered also an infection by CMV. This superimposed infection makes it difficult to link all the CNS sequelae only to ICH (case 5). In our series, another child was found to have CNS sequelae: however, no data exists to allow us to correlate these findings with the presence of ICH at birth. Our data provide evidence to support that the incidence of neurological complications in BP-NAIT is lower than in Zw"-NAIT. However, a larger series will be necessary before one can draw definite conclusions. Therapy of BP-NAIT consists, as in all fetomaternal incompatibilities. in providing antigen-negative platelets, either from the mothers or from phenotyped donors. As expected, the two children in our series who received maternal platelets had a positive and sustained response. The response of 12 neonates receiving random platelets was not clear. Because of the phenotype distribution of BP in the central European populations (about 20%) it is to be anticipated that platelets of four out of five random donors should be compatible. Accordingly, we have seen good responses in a number of cases treated by exchange transfusion and/or random platelets. In cases where repetitive random platelet transfusions were ineffective, other causes for refractoriness (i.e. sepsis, hepatosplenomegaly, insufficient platelet preparation) cannot definitively be excluded. As it has been shown for Zw"-NAIT (Mueller-Eckhardt et al. 1989a), IVIgG is a good therapeutic alternative. The effect of IVIgG could only be properly evaluated in four children in our series. These cases responded favourably. The mode of immunization in BP incompatibility appears somewhat at variance from that established for Zwa immunization. While in the latter condition approximately 50% of the mothers develop Zw" antibodies against fetal antigens already during the first pregnancy, BP antibodies were often detected for the first time in multiparous women in our series. Whether this late antibody recognition reflects 'boostered' sensitization of the mother by successive pregnancies or the
condition did exist previously but only caused mild thrombocytopenia in the previous children, and thus went undiagnosed, remains to be determined. We hypothesize that BP immunization probably arises as early as Zw" immunization, particularly in view of the fact that 'responders' to Zw" and BP immunization have a similar immunogenetic background strong association with HLADR3, DRw52 (Reznikoff-Etievant et al. 1983. 1988: Valentin et al. 1990) or DRw6 (Mueller-Eckhardt et al. 1989b). respectively. We believe that in future prospective studies using sensitive techniques for BP antibody detection, i.e. the MAIPA assay, not only many more 'responders' will be found, but also that the prevalence of asymptomatic Br'NAIT will increase. Because the recurrence of BP-NAIT in families is likely to occur, management of mothers of previously affected children should follow the same rules as for Zw"-NAIT (Kaplan et al. 1988. 1990: Bussel, 1990). We believe that antenatal diagnosis of fetal thrombocytopenia by umbilical cord blood sampling should be recommended in all BP-NAIT mothers. However, we think that interventions should be reserved to cases with a history of severely affected previous children. Concise diagnostic and therapeutic guidelines, however, have yet to be outlined. ACKNOWLEDGMENTS We are grateful to the physicians who provided clinical information concerning their patients. We also thank M. C. Mateos and J. Quesne for the technical assistance and T. Caetano for secretarial assistance. This work was supported in part by Procope 1989-1 990. by the Fondation de France, and by the Deutsche Forschungsgemeinschaft (DFG) (Mu 277/9-7). REFERENCES Bierling, P.. Fromont. P.. Bettaieb. A. & Duedari. N. (1989) Anti-Br" antibodiesin the French population. British lournal of Haematology. 73,428-429.
Bussel, J.B. (1990) Antenatal treatment of neonatal alloimmune thrombocytopenia with intravenous gammaglobulin. Platelet Irnmuno1ogy:fundamental and clinical aspects (ed. by C. Kaplan-Gouet. N. Schlegel. Ch. Salmon and J. McCregor), pp. 287-295. John Libbey Eurotext. Paris. Harrington. W.J.. Sprague. C.C.. Minnich. V., Moore. C.V.. Aulvin. R.C. & Dubach. R. (1953) Immunologic mechanisms in idiopathic and neonatal thrombocytopenicpurpura. Annals of Jnternal Medicine, 38,433-469.
Kaplan. C., Daffos, F., Forestier. F.. Cox, W.L.. Lyon-Caen. D.,DupuyMontbrun. M.C. & Salmon, Ch. (1988) Management of alloimmune thrombocytopenia:antenatal diagnosis and in utero transfusion of maternal platelets. Blood. 72. 340-343. Kaplan C.. Ddos. P.. Forestier, F., Morel, M.C.. Chesnel. N. & Tchemia. G. (1990) Current trends in neonatal alloimmune thrombocytopenia: diagnosis and therapy. Platelet Immunology: fundamental and clinical aspects (ed. by C. Kaplan-Gouet, N. Schlegel, Ch. Salmon and J. McCregor).pp. 267-278. John Libbey Eurotext. Paris. Kiefel. V., Santoso. S.. Weisheit. M. & Mueller-Eckhardt, C. (1987) Monoclonal antibody-specific immobilization of platelet antigens
HPA-Sb Neonatal Alloimmune Thrombocytopenia (MAIPA): a new tool for the identification of platelet-reactive antibodies. Blood. 70, 1722-1 726. Kiefel. V.. Santoso. S.. Katzmann. B. & Mueller-Eckhardt. C. ( 1 988) A new platelet-specific alloantigen Bra. Vow Sanguinis. 54, 101-1 Oh. Kiefel, V.. Santoso. S.. Katzmann. B. & Mueller-Eckhardt. C. (1989) The BP/Brb alloantigen system on human platelets. Blood, 73, 2219-2223. Mueller-Eckhardt, C.. Kiefel. V.. Grubert. A.. Kroll, H.. Weisheit. M.. Schmidt. S.. Mueller-Eckhardt.G. & Santoso, S . ( 1 989a) 348 cases of suspected neonatal alloimmune thrombocytopenia. Lanret, i, 363-366. Mueller-Eckhardt. C., Kiefel, V.. Kroll. H. & Mueller-Eckhardt. G. (1989b) HLA-DRWh. a new immune response marker for immunization agalnst the platelet alloantigen BP. Vox Sanguinis. 57. 90-9 1 . ReznikofT-Etievant. M.F.. Muller. J.Y.. Julien. F. & Patereau. C. (1983) An immune response gene linked to HLA in man. Tissue Antigens. 22. 312-314. ReznikofT-Etievant, M.F.. Kaplan. C.. Muller. J.Y..DafTos. F. & Forestier. F. ( 1988) Alloimmune thrombocytopenias. definition of a group at risk: a prospective study. Current Studies in Hematology arid Blood Transfusion. 55, 119-124.
429
Sharon, R. & Amar. A. (1981) Maternal anti-HLA antibodies and neonatal thrombocytopenia. Lancet. i, 31 3. Shulman. N.R.. Marder. V.J., Hiller, M.C. & Collier, M. (1 964)Platelet and leucocyte iso-antigens and their antibodies: serologic, physiologic and clinical studies. Progress in Hematology (ed. by C. V. Moore and S. B. Brown). pp. 222-304. Grune and Stratton. New York. Soulier. J.P., Patereau. C. & Drouet, J. (1975) Platelet indirect radioactive Coombs test. Its utilisation for PL*’ grouping. Vox Sanguinis, 29, 253-268. Terasaki. P.I.. Bernoco, D.. Park, M.S..Oiturk. G. & Iwaki, Y. ( 1 978) Microdroplet testing for HLA A. B. C and D antigens. American lournal of Clinical Pathology. 69. 1 0 3-1 20. Valentin, N.. Vergracht. A.. Bignon. J.D.. Cheneau. M.L.. Blanchard. D.. Kaplan, C.. Reznikoff-Etievant. M.F. & Muller. J.Y. (1990) HLADRw52a is involved in alloimmunization against PLAi antigen. Human Immunology, 27, 73-79. Von dem Borne. A.E.G.Kr. & Kuijpers. R.W.A.M. (1990) Platelet antigens. new aspects. Platelet Zmmunology:fundamental and clinical aspects (ed.by C. Kaplan-Gouet, N. Schlegel, Ch. Salmon and J. McGregor). pp. 21 9-240. John Libbey Eurotext. Paris.
