Key Points
- •
Fetal platelet disorder is a potentially life-threatening condition.
- •
Fetal thrombocytopenia may lead to fetal bleeding. complications, such as an intracranial haemorrhage.
- •
Idiopathic thrombocytopenic purpura.
- •
Has an incidence of 1 to 2 in 1000 pregnancies.
- •
Causes severe fetal thrombocytopenia in 5% to 20% of the cases.
- •
Rarely leads to bleeding problems in fetuses or neonates.
- •
Is treated primarily with corticosteroids.
- •
- •
Fetal and neonatal alloimmune thrombocytopenia.
- •
Occurs in 1 in 1000 pregnancies.
- •
Is mainly caused by human platelet antigens 1a (80%) and 5b (10%) in Caucasians.
- •
Causes severe bleeding complications in 10% of cases of severe thrombocytopenia.
- •
Treatment should be noninvasive with intravenous immunoglobulins.
- •
Population-based screening would mean a major improvement in the prevention of bleeding complications.
- •
Introduction
Fetal platelet disorders, causing fetal thrombocytopenia, are relatively rare but potentially life-threatening conditions. During normal fetal life, the platelet count progressively increases, and reaches a level of approximately 150 × 10 9 /L by the end of the first trimester. The normal range for platelet counts in healthy fetuses and neonates is equal to that of adults (150–450 × 10 9 /L). Therefore fetal and neonatal thrombocytopenia is defined as a platelet count less than 150 × 10 9 /L regardless of gestational age, which corresponds with values below the fifth percentile, calculated in adults. The degree of the thrombocytopenia can be further classified to mild (100–150 × 10 9 /L), moderate (50–100 × 10 9 /L) or severe (<50 × 10 9 /L). In contrast to neonatal thrombocytopenia, the exact frequency of fetal thrombocytopenia is unknown. Of all newborns, 1% to 2% have a platelet count below 150 × 10 9 /L and 1 to 2 of 1000 newborns have a severe thrombocytopenia.
The risk for fetal thrombocytopenia is the development of bleeding complications, varying from harmless skin bleeds to internal organ haemorrhage, intracranial haemorrhage (ICH) or even perinatal demise. These complications mostly present after birth, and a diagnosis of a fetal platelet disorder occurs postnatally in the vast majority of cases. Therefore preventive measures can only be taken in subsequent pregnancies ( Table 41.1 ).
| Increased destruction |
|---|
| Immune thrombocytopenia |
|
|
|
|
| Peripheral consumption |
|
|
|
|
| Decreased production |
|---|
| Genetic disorders (TAR syndrome, trisomy 13,18,21, triploidy, Turner syndrome, amegakaryocytosis, Wiskott-Aldrich syndrome, May-Hegglin syndrome, Bernard-Soulier syndrome, Alport syndrome) |
| Bacterial infection (GBS, E scherichia c oli , Listeria spp., syphilis) |
| Viral infection (CMV, parvovirus, rubella, HIV, HSV) |
| Parasite infection (toxoplasmosis) |
| Asphyxia |
| Placental insufficiency (preeclampsia, IUGR, diabetes, premature birth) |
Nonimmune Causes for Fetal or Neonatal Thrombocytopenia
Nonimmune conditions that are associated with fetal and neonatal thrombocytopenia act through increased destruction of platelets as well as a decreased production. Non–immune-mediated increased destruction or consumption can be caused by disseminated intravascular coagulation (DIC), thrombosis, Kasabach-Merritt Syndrome or hypersplenism. Through increased destruction, placental insufficiency (premature birth, preeclampsia, intrauterine growth restriction, diabetes), several genetic abnormalities, infection and asphyxia can lead to fetal and neonatal thrombocytopenia.
Immune Causes of Fetal and Neonatal Thrombocytopenia
Because immune-mediated fetal platelet disorders are the most important cause of severe fetal thrombocytopenia, responsible for one third of all neonatal thrombocytopenia cases, this chapter focuses on idiopathic thrombocytopenic purpura (ITP) and fetal and neonatal alloimmune thrombocytopenia (FNAIT). Also, through an unknown mechanism, a small proportion of cases with severe fetal anaemia caused by red blood cell (RBC) alloimmunisation is associated with fetal thrombocytopenia.
Autoimmune or Idiopathic Thrombocytopenic Purpura
Maternal thrombocytopenia is encountered regularly, complicating 1 in 12 pregnancies. The most common cause of maternal thrombocytopenia is a benign transient condition called gestational thrombocytopenia, accounting for approximately two thirds of all cases of maternal thrombocytopenia; ITP accounts for 3%. Other causes of maternal thrombocytopenia are preeclampsia, HIV, systemic lupus erythaematosus and thyroid dysfunction.
There are two different types of ITP: the acute form and the chronic form. Acute ITP is predominantly a condition of childhood that seldom occurs in pregnancy. It is mostly preceded by a viral infection and is caused by cross-reactivity between viral antigens and platelet antigens. This form usually resolves within weeks or months. ITP in pregnancy is therefore almost always chronic ITP, with an incidence of 1 to 2 in 1000 pregnant women.
Pathophysiology
Chronic ITP is an autoimmune disorder caused by the maternal production of antibodies against glycoproteins present on the membranes of maternal platelets. The majority of these autoantibodies are of the immunoglobulin (Ig) G class and are thus able to cross the placental barrier, bind to fetal platelets and cause fetal thrombocytopenia. The reported incidence of severe neonatal thrombocytopenia in ITP varies between the 5% and 20%. Although these numbers vary widely among studies, none of them reported any significant bleeding problems in neonates as a result of the ITP. No cases of severe in utero bleeding have been documented, and the reported incidence of ICH is low and varies between 0% and 1.2%. The lowest platelet count in the affected newborns mostly occurs within 7 days after birth.
In pregnancy, maternal platelet counts are usually lower than before. Maternal symptoms can vary from none to severe haemorrhaging but are usually mild, and pregnant women seem to have a greater tolerance to ITP compared with nonpregnant women because of the procoagulant state in pregnancy.
Unfortunately, no reliable pregnancy-specific parameters exist to predict the severity of fetal thrombocytopenia in ITP. The maternal platelet count, nor IgG level seems to correlate with the fetal platelet count. The strongest correlation found so far is the lowest platelet count in older siblings. The only maternal factor identified to predict a low platelet count in affected neonates is a history of splenectomy.
Diagnosis
The diagnosis of ITP is one of exclusion, other causes of thrombocytopenia during pregnancy (e.g., preeclampsia, gestational thrombocytopenia, HELLP (haemolysis, elevated liver enzymes, low platelet count) syndrome, DIC, massive obstetric haemorrhage or acute fatty liver syndrome) need to be ruled out first. The distinction between gestational thrombocytopenia and the first presentation of ITP may be particularly difficult. Differences are the gestational age at detection and platelet count. ITP is usually detected in early pregnancy because the condition is present before conception, and it usually has a lower platelet count than gestational thrombocytopenia. In addition, an elevated mean platelet volume, implicating increased platelet production, can support the diagnosis. Bone marrow examination is not performed during pregnancy but would reveal normal or elevated megakaryocytic numbers. Thrombopoietin (Tpo) level can be used to distinguish platelet production disorders from platelet destruction disorders. In ITP, Tpo levels are normal or slightly elevated in contrast to significant elevation of Tpo levels in patients with platelet production disorders.
Autoantibodies can be identified using a platelet immunofluorescence test (PIFT), which is based on the detection of platelet-bound antibodies. Unfortunately, it has a relatively low sensitivity of 70% in patients with ITP. Positive reactions can be identified in case of greater than 1000 molecules of IgG bound to a platelet; 450 molecules can already cause platelet destruction. Also, in some patients with ITP, destruction of platelet can be T cell mediated, so platelet-bound antibodies will not be measurable.
Obstetric Management
Prepregnancy
Idiopathic thrombocytopenic purpura is not a reason to discourage pregnancy; however, women with severe thrombocytopenia despite a splenectomy and high doses of corticosteroids are at high risk for complications and should have extensive preconception counselling about possible risks. When planning pregnancy, it is wise to optimise treatment before conception, mainly to assess the need for splenectomy before pregnancy.
Prenatal
Management of IPT during pregnancy requires a multidisciplinary approach, with a team including a haematologist, paediatrician (neonatologist), obstetrician and anaesthetist. A complete overview is displayed in Fig. 41.1 .
During pregnancy, there are three indications for starting treatment: symptomatic women, a platelet count below 30 × 10 9 /L or the need for a higher platelet count before a procedure (e.g., caesarean section).
When platelet counts are higher than 30 × 10 9 /L, no treatment is necessary, and monitoring of maternal platelet counts needs to be performed every 2 weeks and more closely as delivery approaches.
First choice of treatment during pregnancy is similar to the approach in nonpregnant women, prednisone started at a dosage of 1 to 2 mg/kg/day and then tapered to find the minimal effective dose. If the thrombocytopenia shows to be resistant to corticosteroid treatment with prednisone or in case of serious side effects, intravenous infusion of immunoglobulins (IVIG) is next in line. Other treatments with limited evidence for their efficiency in ITP during pregnancy are intravenous anti-D infusion, splenectomy and azathioprine. Rituximab, danazol, Tpo receptor agonists and most other immunosuppressive drugs should not be administered during pregnancy because of possible teratogenicity. Platelet transfusions are only to be administered in very severe thrombocytopenia (platelet counts <20 × 10 9 /L) and at times when there is a high risk for bleeding.
Labour and Delivery
Maternal platelet counts above 50 × 10 9 /L are considered to be safe for vaginal as well as caesarean delivery. Therefore, if predelivery platelet counts are below 50 × 10 9 /L, IVIG treatment should be administered at a dose of 0.8 kg/day. Although guidelines can differ among countries and centres, a platelet count above 80 × 10 9 /L is considered safe for epidural analgesia. Treatment during pregnancy does not seem to have any effect on fetal platelet counts or the occurrence of neonatal bleeding problems; it should be administered for maternal indications only.
Initially, the management of labour and delivery in ITP patients was based on the concerns of ICH in the thrombocytopenic neonate, secondary to vaginal birth trauma. Therefore fetal scalp blood sampling or cordocentesis was performed, and caesarean delivery was preferred when fetal platelet count was below 50 × 10 9 /L. Fetal scalp blood sampling appeared to yield falsely low counts, and predelivery cordocentesis, although producing reliable fetal platelet counts, has a significant risk for fetal loss and severe morbidity. In combination with the low incidence of perinatal bleeding, these should not be performed as routine procedures in patients with ITP. The route of delivery (vaginal or caesarean section) does not seem to affect the incidence of ICH; therefore, caesarean section in patients with ITP should only be performed for obstetric indications. Interventions such as fetal scalp blood sampling and vacuum extraction should be avoided during delivery.
Postnatal
An initial blood sample should be taken from the umbilical cord to assess the neonatal platelet count. Second, daily neonatal platelet counts should be done for approximately 1 week, depending on the presence and degree of neonatal thrombocytopenia detected. In addition, haemoglobin and bilirubin counts should be monitored. Cranial ultrasound is indicated only in case of severe thrombocytopenia. Postnatal treatment with IVIG is only indicated in case of severe neonatal thrombocytopenia (<50 × 10 9 /L), and platelet transfusion should only be used in case of neonatal bleeding. Breastfeeding is not contraindicated.
Fetal and Neonatal Alloimmune Thrombocytopenia
Fetal and neonatal alloimmune thrombocytopenia can originate after maternal alloimmunisation against paternally derived human platelet antigens (HPAs) on fetal platelets. FNAIT is the main cause of thrombocytopenia in newborns. Approximately half of the severe neonatal thrombocytopenia cases (platelet counts <50 × 10 9 /L) result from FNAIT.
Natural History
The maternally produced alloantibodies in FNAIT are of the IgG subclass and are thought to enter the fetal circulation by crossing the placental barrier using the neonatal Fc receptor (FcRn) and, after entering, cause fetal thrombocytopenia. This thrombocytopenia can be completely asymptomatic or lead to a wide spectrum of fetal and neonatal bleeding problems. These bleeding problems can vary from relatively harmless skin bleeding, such as petechiae, purpura or haematoma, to more severe haemorrhage ( Fig. 41.2 ). FNAIT is associated with a high risk for neonatal mortality or morbidity, with many surviving children with neurodevelopmental impairments, such as bilateral blindness, cerebral palsy or delayed cognitive development. An ICH is the most serious bleeding complication. The majority of these ICHs affect the first-born child, an important difference to RBC alloimmunisation, which does not normally affect the first child. A multicentre study that reviewed the largest series of ICHs reported that 54% of the bleedings occurred before the 28th week of gestational age, and 67% started before 34 weeks’ gestational age. Although a lot of research is being performed on identifying risk factors for developing ICH, the only useful predictor of ICH thus far is a sibling that had ICH. Based on retrospective data on 33 untreated successive pregnancies, the recurrence rate is estimated at 79%. In addition to ICH, relatively unknown massive internal organ haemorrhage, such as pulmonary or gastrointestinal bleeding, seem to occur as well and are associated with a considerable risk for neonatal death.
Pathophysiology
Platelet Alloantigens and Alloantibodies
The platelet specific alloantigens, or HPAs, are carried by glycoprotein complexes that are present on the platelet membrane. Different HPA epitopes are created by single nucleotide polymorphisms (SNPs) that result in small changes in the glycoprotein structure through an amino acid change. Currently, there are 37 different types of HPAs described, located on six different glycoprotein complexes (IIb/IIIa, Ib/IX, Ia/IIa and CD109) on the platelet membrane.
For a recent overview, see http://www.ebi.ac.uk/ipd/hpa/table1.html . The 12 most frequently involved HPAs are clustered into six biallelic groups ( Table 41.2 ). The greater portion of all HPA are localised on glycoprotein (GP) IIb/GP IIIa, also called integrin αIIbβ3, the fibrinogen receptor, which is the most abundant molecule on the surface of platelets. GP IIb contains 7 and GP IIA 16 different HPA systems. Although all HPAs that are present on GP IIIa can be involved in FNAIT, HPA-1a is the predominant cause of FNAIT, especially in the Caucasian population. HPA-1a accounts for approximately 80% of cases followed by HPA-5b, responsible for about 10% of cases of FNAIT ( Table 41.3 ). In the Asian population, however, anti–HPA-4b is the most frequently involved alloantibody followed by anti–HPA-3a and anti–HPA-21b. Furthermore, alloantibodies against GP IV/CD36 seem to have an increased presence in the African and Asian populations.
| Antigen | Original names | Glycoprotein | CD | Nucleotide change | Amino acid change | Frequency (%) | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| African | Allele | Asian | Genotype White | |||||||
| HPA-1a | Zw a , Pl A1 | GP IIIa | CD61 | T176 | Leu 33 | 1A | 90 | 100 | 1a/a | 72 |
| HPA-1b | Zw b , Pl A2 | GP IIIa | C176 | Pro 33 | 1B | 10 | 0 | 1a/b | 26 | |
| 1b/b | 2 | |||||||||
| HPA-2a | Ko b | GP Ibα | CD42b | C482 | Thr 145 | 2A | 71 | 95.2 | 2a/a | 85 |
| HPA-2b | Ko a , Sib a | GP Ibα | T482 | Met 145 | 2B | 29 | 4.8 | 2a/b | 14 | |
| 2b/b | 1 | |||||||||
| HPA-3a | Bak a , Lek a | GP IIb | CD41 | T2621 | Ile 843 | 3A | 68 | 59.5 | 3a/a | 37 |
| HPA-3b | Bak b | GP IIb | G2621 | Ser 843 | 3B | 32 | 40.5 | 3a/b | 48 | |
| 3b/b | 15 | |||||||||
| HPA-4a | Yuk b , Pen a | GP IIIa | CD61 | G506 | Arg 143 | 4A | 100 | 99.5 | 4a/a | >99 |
| HPA-4b | Yuk a , Pen b | GP IIIa | A506 | Gln 143 | 4B | 0 | 0.5 | 4a/b | <0.1 | |
| 4b/b | <0.1 | |||||||||
| HPA-5a | Br b , Zav b | GP Ia | CD49b | G1600 | Glu 505 | 5A | 82 | 98.6 | 5a/a | 88 |
| HPA-5b | Br a , Zav a , Hc a | GP Ia | A1600 | Lys 505 | 5B | 18 | 0.4 | 5a/b | 20 | |
| 5b/b | 1 | |||||||||
| HPA-15a | Gov b | CD109 | CD109 | C2108 | Ser 703 | 15A | 65 | 53.2 | 15a/a | 35 |
| HPA-15b | Gov a | CD109 | A2108 | Tyr 703 | 15B | 35 | 46.8 | 15a/b | 42 | |
| Authors | P atients ( n ) | Alloantibody detected | Frequency (%) | Alloantibody detected | Frequency (%) |
|---|---|---|---|---|---|
| Mueller-Eckhart et al. (1989) | 106 | Anti–HPA-1a Anti–HPA-5b Anti–HPA-1b | 90 8 0.8 | Anti-HPA-3a Anti HPA-1a + 5b Anti-B | 0.8 0.8 0.8 |
| Porcelijn (2004) | 217 | Anti–HPA-1a Anti–HPA-5b Anti–HPA-3a Anti–Priv Ag | 73.7 14.7 4.6 1.5 | Anti HPA-1b Anti HPA-15a Anti HPA-15b Anti A or anti B | 1.4 0.5 0.5 2.8 |
| Davoren et al. (2004) | 1162 | Anti–HPA-1a Anti–HPA-5b Anti–HPA-1b Anti–HPA-3a Anti–HPA-5a Anti–HPA-3b | 79 9 4 2 1 0.8 | Anti GPIV (CD36) Anti HPA-4a Anti HPA-4b Anti HPA-6bw Combinations | 0.4 0.1 0.1 0.1 3.1 |
| Knight et al. (2011) | 151 | Anti–HPA-1a/b Anti–HPA-5a/b | 81 7 | Anti HPA-1a + 5b Other | 5 7 |
The immune response to alloantigens is mediated by human leukocyte antigen (HLA) class II molecules. Antigen-presenting cells such as dendritic cells, macrophages and B lymphocytes process the antigens into small peptides, which are presented by the HLA class II antigens on their surfaces. CD4-positive T-helper lymphocytes can recognise the peptide–HLA complex, leading to the activation of the T cell. Activated T cells then interact with B lymphocytes, initiating antibody production. After active transport, through FcRn, of the formed alloantibodies against the fetal antigens enter the fetal circulation and induce destruction of fetal platelets.
The GP that carries the most important alloantigen in FNAIT, HPA-1a, is GP IIIa, which is also called ‘integrin β3’. On platelets, integrin β3 is complexed to GP IIb. In addition to this complex, integrin β3 can form a complex with αV as well. This αVβ3 complex, still carrying the HPA-1a epitope, is scarcely expressed on platelets but is prominently present on the surface of endothelial cells, vascular smooth muscle cells and syncytiotrophoblast cells. The presence of HPA-1a on these different cell types, other than platelets, and interaction of maternal alloantibodies with these cells might reveal new mechanisms in the pathophysiology.
Vascular Integrity
The exact pathogenic mechanism leading to devastating ICH in FNAIT has never been adequately understood. Generally, alloantibodies were thought to enter the fetal circulation and cause bleeding problems through destruction of fetal platelets resulting in thrombocytopenia. However, only a small proportion of the severely thrombocytopenic newborns have a severe haemorrhage, indicating another mechanism to be involved. Because integrin β3 that carries the HPA-1a epitope is also present on endothelial cells (in complex with αV; αVβ3), it has been suggested that interaction of anti–HPA-1a with endothelial cells plays a key role in the development of bleeding complications. In vitro studies already illuminated the direct interaction between anti–HPA-1a and human umbilical vein endothelial cells (HUVECs), demonstrated by a decreased HUVEC spreading as well as a decreased integrity of their monolayer in electric cell–substrate impedance sensing (ECIS) assays. In addition, in vivo animal studies showed that mice without circulating platelets and or fibrinogen do not show any bleeding problems in utero . This supports the hypothesis that instead of the thrombocytopenia and blood coagulation, another mechanism plays a key role in causing bleeding complications. Recently, a large study with both active and passive murine models of anti–HPA-1a mediated FNAIT has been performed and showed that ICHs in these mice occurred regardless of platelet count. Also, HPA-1a antibodies inhibited angiogenic signalling, induced endothelial cell apoptosis and decreased vessel density in affected brains as well as retinas. This delivered the first direct evidence for αVβ3 integrin’s role in angiogenesis and HPA-1a–induced impaired angiogenesis in mice, suggesting this to be a key factor in causing bleeding complications in FNAIT. The first analysis with a small number of human sera containing HPA-1a antibodies concluded that the interaction of the antibodies with endothelial cells might determine and possibly predict the occurrence of ICH.
Placental Function
In addition to platelets and endothelial cells, integrin β3 is also expressed on placental tissue by syncytiotrophoblast cells in αVβ3 complex as well. Although there is no direct evidence, it has been suggested that anti–HPA-1a might induce placental insufficiency through interaction with these syncytiotrophoblast cells, demonstrated by a strong association with intrauterine growth restriction, as well as cases of intrauterine fetal demise (in absence of bleeding problems). In addition, in a group of 13 FNAIT cases, lymphoplasmacytic chronic villitis was found to be significantly more present in placenta tissue of untreated cases when compared to cases treated with IVIG. Another correlation that has been suggested is one between FNAIT and miscarriages. In addition, the expression of HPA-1a on placental tissue might lead to increased and early exposure to the fetal HPA-1a and might be a possible explanation for FNAIT already affecting first pregnancies and first-born children in FNAIT.
Overall, there is increasing evidence of more advanced pathophysiologic mechanisms initiated by anti-HPA-1a, than solely platelet destruction.
Human Leukocyte Antigen
Specific HLA class II molecules seem to be associated with the occurrence of alloimmunisation in HPA incompatible pregnancies. HLA-DRB3∗0101 was found to be positively correlated to the occurrence HPA-1a alloimmunisation in HPA-1a–negative pregnant women. The underlying mechanism suggested is the difference in the ability to form stable bonds between Pro33/Leu33, the amino acids present on the integrin β3 peptides of HPA-1b alleles and HPA-1a alleles, respectively.
There is some evidence that HLA class I plays a role in FNAIT as well. In contrast to HLA class II molecules, HLA class I molecules are expressed on platelet membranes. With an expression of approximately 20,000 HLA class I molecules, platelets are the major source of HLA class I present in blood. Maternal alloimmunisation against fetal HLA on platelets, possibly causing antibody-mediated fetal thrombocytopenia in a similar mechanism to HPA alloantibodies is mentioned in a number of case reports. In contrast to what previously has been thought, HLA class I alloantibodies can cross the placental barrier and do enter the fetal circulation. Whether or not these antibodies can actually be held responsible for fetal and neonatal thrombocytopenia is not clear yet. Strong evidence to support such a causal relationship is currently lacking.
Incidence
In the absence of population-based screening programs, estimates of incidence of FNAIT are mainly based on retrospective data and some small prospective cohort studies and therefore vary widely. It is generally accepted that FNAIT is likely to be underdiagnosed, with only 37% of cases of severe FNAIT detected in the absence of screening programs. An Irish cohort study estimated that only 7% of expected cases are detected clinically. Based on a systematic review pooling results on postnatal screening of platelet counts in 59,425 newborns, the incidence of severe FNAIT is 0.04% (1 in 2500 newborns), leading to an ICH in 25% of these (1 in 10,000).
Focussing on the predominant cause of FNAIT, HPA-1a alloimmunisation, a few prospective studies have attempted to estimate its incidence. The largest prospective screening study, including 100,488 pregnant women in Norway, reported an incidence of HPA-1a–negative women of 2.1%, leading to HPA-1a alloimmunisation in 10.7%. All alloimmunised women underwent an elective caesarean section at 2 to 4 weeks before term, so neonatal platelet counts and the incidence of bleeding problems are potentially an underestimation. They reported 58% of alloimmunisations to result in FNAIT, 33% in severe FNAIT and 2% of all alloimmunisations developed an ICH. Another, much smaller, prospective screening study reported an overall incidence of FNAIT caused by HPA-1a of 1 in 1163 live births and an incidence of severe FNAIT of 1 in 1695. The most accurate estimation of the population incidence of FNAIT has been made by Kamphuis and associates, who performed a systematic review including prospective studies on HPA-1a screening. They concluded that HPA-1a alloimmunisation occurred in 9.7% of pregnancies at risk, leading to severe FNAIT in 31% of the cases and to perinatal ICH in 10% of the severe FNAIT cases ( Fig. 41.3 ).
